Physical AI Brief

On-policy distillation (OPD) grounds token-level supervision in the student's own trajectory, yet suffers from prefix failure: once the student commits to a wrong reasoning direction, all subsequent generation builds on this deviation, producing misdirected continuations that elicit unreliable supervision and waste compute. We identify a teacher-student continuation asymmetry on failed prefixes, where the teacher tends to redirect while the student continues along the original direction, and convert it into a label-free handoff trigger in Relay On-Policy Distillation (Relay-OPD). During training, Relay-OPD constructs relay trajectories by letting the teacher briefly take over at detected trigger points to produce a teacher leg, after which the student resumes and is optimized on the resulting trajectory. A limited relay budget concentrates intervention on critical early positions while limiting departure from the student policy. With a Qwen3-4B-Instruct-2507 teacher and Qwen3-0.6B/1.7B-Non-Thinking students on eight mathematical reasoning benchmarks, Relay-OPD achieves the best or second-best results on every benchmark, outperforming standard OPD by +5.73% and the strongest baseline FastOPD by +1.49% on average for 1.7B, with consistent gains at 0.6B. Training trajectory length is reduced by over 50%.

Forward latent world models predict how actions change a scene, but recover actions for a desired change only through expensive test-time search. We introduce INTACT (INtent-To-ACTion), an end-to-end JEPA that turns action-labeled, reward-free trajectories into a deployable intent-to-action interface. Each transition supplies physical intent $z_{t+1}-z_t$, while a future goal supplies deployment intent $\operatorname{sg}(z_g)-z_t$. The architecture is isomorphic between the local and goal motion-intent backbone-input graphs through an identical four-slot grammar and shared parameters, and between supported local and goal motion-intent families through action-law semantics induced by the same predictor rather than pointwise latent equality. INTACT also provides intact transfer from RGB evidence to action-effective latent intent coordinates and from intent families to their corresponding action-law families. Asymmetric endpoint gradients ground physical successors and fix future goals as anchors, joining representation learning and control without pointwise latent matching or globally linear dynamics. The resulting coordinates support a robust distributional action law: its conditional mean serves directly as a search-free policy, while sampling remains available for diversity or optional verification. On the four official LeWM tasks, one-epoch, zero-search models reach 85.78\%, 100.00\%, 97.67\%, and 97.89\% success. Optional local CEM centered on the Direct plan reaches 96.86\% macro success using 384 instead of 9,000 candidate sequences, reducing sampling by $23.44\times$ while improving pure CEM by 16.00 points. One shared four-task encoder reaches 89.39\% E5 Direct macro and improves every task over jointly trained LeWM, while predicted--expert action-family kNN tracks Direct success at $r=0.954$. Direct inference takes 2.9--5.5 ms.

Generalist manipulation policies increasingly take the form of action-chunking flow policies built on large pretrained backbones. Such chunks run open-loop, so the policy cannot react to sensory input arriving mid-execution, sacrificing \emph{reactivity}. Replanning more often would restore it, but the perception-to-action pipeline (a large backbone plus multiple denoising steps) is too slow: this \emph{latency} forbids frequent replanning and leaves committed actions stale, making such policies ill-suited for dynamic, closed-loop control. We present $π\mathbf{R}^2$, which makes these policies reactive and real-time while retaining large backbones, expressive multi-modal policies, and multi-action prediction. Built on the per-position noise schedule of diffusion forcing, $π\mathbf{R}^2$ contributes two ideas. First, it splits conditioning into a fast channel (proprioception, fresh every tick) and an asynchronously updated slow channel (vision-language features), so the policy reacts to proprioception within a chunk while tolerating stale vision. Second, a latency-adaptive flow schedule treats in-flight actions as inpainting conditioning and emits actions in one denoising step per call, letting one trained model adapt to varying hardware latency. Requiring minimal modification to existing architectures, $π\mathbf{R}^2$ can be finetuned from a pretrained policy: applied to GR00T-N1.7 on a real xArm6+XHand platform, it replans closed-loop roughly $4\times$ faster than the base policy (~$25$Hz on an A5000 GPU), acting on a fresh observation every $40$ms. Across simulation and real-world manipulation tasks, $π\mathbf{R}^2$ improves the success rate by up to $23\%$ in simulation and $30\%$ in the real world over the strongest baseline. Project page: https://pi-r2-flow.github.io/

Mixture-of-Experts (MoE) variants of Low-Rank Adaptation (LoRA) route every token to a fixed number of experts $k$. Tokens differ in how uncertain the model is about them, so a single k over-spends on easy tokens and under-serves hard ones. We observe that the router's output distribution is already a per-token uncertainty signal: peaked mass indicates confidence, while a flat distribution indicates ambiguity. We introduce CARE (Confidence-Adaptive Routing of Experts), which admits experts in a nucleus fashion. Experts are activated in decreasing router weight until their cumulative mass reaches a threshold, with a small extension when the admitted experts disagree. A budget thermostat calibrates the threshold so that the average number of active experts matches any target. CARE is a drop-in, single-forward-pass rule with no extra parameters. Across eight commonsense benchmarks on LLaMA-3.1-8B and Qwen2.5-7B, as well as math, code, and knowledge tasks, CARE improves over fixed top-k MoE-LoRA at matched compute and matches the fixed-k=4 baseline while activating fewer experts. The same confidence and disagreement signals also improve out-of-distribution detection over MSP, entropy, and multi-pass proxies. We support the design with nucleus fidelity, budget optimality, and an epistemic reading of disagreement, and we release code.

Acoustic information provides rich cues about object location, material properties, and changes caused by contact or motion. This paper introduces a new set of acoustic-aware manipulation tasks for imitation learning, in which robots must use auditory cues to determine manipulation targets. These tasks require sound source localization and identification for active exploration in robotic manipulation. Also, we propose a multimodal imitation learning framework, Spatial-Spectral Audio Action (S2A2), that integrates visual features with acoustic spatial and acoustic signal information for the acoustic-aware manipulation tasks. We implemented S2A2 models that integrates policies such as ACT, Diffusion Policy, VQ-BeT, and $π_0$, into our framework. Simulation experiments showed that the proposed method is the most effective for tasks requiring both position and timbre. Furthermore, real-robot experiments confirm the applicability of the proposed tasks and framework to real-world manipulation.

The modeling of photonuclear reactions is increasingly important for applications involving high-energy photon fields, including accelerator-driven neutron sources, radiation shielding, medical physics, and fusion technologies. Although the Monte Carlo code OpenMC provides well-verified photoatomic transport capabilities, its photonuclear physics is currently available only in an unofficial development branch and requires independent verification before broader scientific use or possible integration into the official code distribution. This work presents a systematic verification of the OpenMC photonuclear implementation using six single-collision broomstick benchmarks based on 2H, 9Be, and 238U targets irradiated by monoenergetic 5 MeV and 15 MeV photons and by a continuous 1--20 MeV linear accelerator (LINAC)-representative spectrum. OpenMC was compared with MCNPX using common ENDF7u photonuclear data, with FLUKA using its native photonuclear models, and with a first-collision analytical solution for the integrated neutron yield. Calculations using the IAEA/PD-2019 library were also performed to quantify nuclear-data sensitivity. OpenMC and MCNPX agreed within 0.7% in integrated neutron yield for all benchmark cases when the same ENDF7u data were used. FLUKA, which relies on its own native photonuclear models rather than ENDF7u, differed from the analytical solution by approximately 6-9% for the monoenergetic cases and by no more than approximately 4% for the continuous-source cases. Changing the OpenMC library to IAEA/PD-2019 produced deviations of up to 11.3% from the ENDF7u-based analytical solution, with the sensitivity varying strongly by nuclide and source spectrum.

We present VetClaw, an edge-cloud multimodal agentic system for early veterinary disease screening. VetClaw uses a camera module as an edge sensing device and sends captured images, together with optional symptom descriptions, to a server-hosted vision-language model for zero-shot disease classification. The system separates agent interaction from workflow orchestration: OpenClaw provides scheduling, tool access, user interaction, and notification services on the edge device, while LangGraph manages the stateful screening workflow, including input validation, image transmission, model invocation, safety checks, conditional routing, failure handling, and structured logging. This design moves beyond static image classification by enabling the system to collect visual evidence, invoke external models, apply deterministic safety rules, and generate diagnostic-support alerts. Results show that image-only VLM prediction remains limited, whereas symptom-guided and multimodal inputs improve zero-shot classification performance. Thus, VetClaw transforms a static prediction model into a coordinated, safety-aware system that can use tools, manage workflows, handle failures, and escalate uncertain cases.

Computer-use agents (CUAs) increasingly act through desktop GUIs to complete long-horizon tasks. Current benchmarks primarily measure end-task success or single-frame grounding. Neither isolates whether a model can reconstruct the causal, task-relevant transition produced by an action- crucial for rejecting stale observations, verifying progress, and recovering from failure. This is difficult because inference, remote input, app rendering, and screenshot capture are asynchronous: the next observation may be delayed, occluded, transient, or unrelated, then misread as progress and carried into subsequent planning. We introduce Desktop-Delta Bench (DDB), an offline step-level benchmark with 2,013 human-verified instances from novel, multi-app Linux trajectories across ~15 applications and 50 task domains. DDB trajectories targets 3 failure dimensions- state verification, source tracking, and context-aware control- through 2 complementary tasks: 463 3-frame temporal-ordering instances, including 105 with a cross-trajectory decoy, and 1,550 before-after pairs labeled from 5 actions + its payload. We evaluate 8 closed and open-source model families across 32 ordering and 16 single-action settings, observing consistent gaps. Ordering remains unsaturated: best non-decoy and decoy exact-match rates are 65.1% and 65.7%. Task context improves decoy identification by 6.9 percentage points but reduces non-decoy exact match by 2.2 points; error analysis reveals systematic copying of the presented A-B-C order. Single-action results show that inferring the action family is harder than locating it: click F1 is 0.96 vs, 0.76 for drag, while recognized drags are generally localized well. DDB, thus, complements end-to-end benchmarks by filling the missing diagnostic layer between GUI grounding and final task success, enabling targeted improvements to desktop CUA verification, reliability, and recovery.

Much like humans benefit from guidance while learning, reinforcement learning algorithms may benefit from additional supervision beyond rewards. Leveraging additional information during training to learn better representations and behaviors has been the focus of asymmetric reinforcement learning. This learning paradigm has proven effective under partial observability when additional state information is available, but also under full observability when more refined state information is available. Focusing on model-based reinforcement learning, we study the effect of asymmetric learning on observation representations and on privileged information representations. First, we identify a limitation in the privileged information representations learned by an asymmetric model-based algorithm known as the Informed Dreamer. Then, we propose a novel asymmetric representation learning objective using latent guidance, resulting in a new algorithm called the Reinformed Dreamer. Experiments across several benchmarks show a more consistent improvement over Dreamer than previous asymmetric approaches.

We present Wonder, a general-purpose video world model for real-time, camera-controllable world exploration. Given an image or a conditional video, Wonder constructs a playable world where users can navigate interactively by moving the camera, discovering unseen regions, and revisiting previously observed areas in real time and over a long-term horizon. Achieving this capability requires a system-level co-design of control method, memory mechanism, and training strategy. We introduce a novel camera conditioning with a dense coordinate field whose renderings provide spatially aligned motion and orientation cues, allowing the model to interpret camera motion directly as visual evidence. To support fast and precise memory retrieval over a growing generation context, we propose an efficient sparse attention-based memory mechanism, enabling the model to selectively attend to a small set of relevant context tokens at inference time, regardless of actual context length. We further develop several techniques to rectify the self-forcing-style distillation pipeline, improving the student model's ability to respect control signals, as well as maintaining diverse generation modes and long-term memory from the teacher. Together, these components enable Wonder to synthesize diverse, minute-scale videos at 16 FPS while preserving coherent geometry, appearance, and dynamics across long rollouts. Beyond image-to-video generation, Wonder naturally supports video-conditioned generation, allowing existing dynamic scenes to be re-shot in real time.

Memory is essential for LLM agents to accumulate task experience and reuse task-specific execution strategies. However, real-world deployment over boundary-agnostic and evolving task streams exposes a fundamental stability-plasticity dilemma. External retrieval-based memory can rapidly absorb new evidence, but it often fails to internalize recurring execution patterns and incurs inference-time retrieval overhead. Parametric memory enables stable and efficient execution once learned, but typically relies on explicit task boundaries and fixed parameter budgets. Inspired by the human brain, which balances plasticity and stability through complementary episodic storage and gradual consolidation, we propose UniMem, a self-routing framework for autonomous memory management. UniMem uses learnable routing tokens as memory controllers, enabling adaptive coordination between complementary memory pathways: novel or sparse tasks are retained in an episodic buffer for retrieval-augmented execution, while recurring and reliable patterns are consolidated into expandable parametric memory. By decoupling task identification from task execution with routing tokens and parametric memory blocks, UniMem expands memory on demand without task labels during deployment or uncontrolled parameter growth. Experiments on long-horizon streaming task sequences show that UniMem consistently outperforms baselines while maintaining execution fidelity, achieving an average gain of 4.0 EM points across three backbone models.

Recently, photonic transformer accelerators (PTAs) have successfully achieved significant speedup and energy efficiency improvements over electronic accelerators for expediting Transformer inference. However, state-of-the-art rely on expensive multi-wavelength light generation and large dot-product units due to active phase-shifter components, thus making their approach inefficient and impractical. To address this, we propose MDTransformer, a novel hardware-software co-design of PTA based on mode-division optical dataflow and operations. Specifically, MDTransformer performs complex matrix operations using spatial-mode interference, that leverages the inverse-designed multi-mode couplers, crossings, and Mach-Zehnder IQ modulators into a compact mode-division photonic tensor core (MPTC), capable of executing matrix multiplications in the optical domain. Its each guided mode (i.e., TE0-TE3) acts as an independent computational lane, enabling four-fold parallelism-per-waveguide without spectral filtering or free-spectral-range limitations. Moreover, its coherent detection and IQ modulation jointly encode amplitude and phase, realizing complex-valued arithmetic for full-range operations in transformers. MDTransformer offers analog multiplication with sub-4-bit effective precision and inter-modal crosstalk below -30 dB. Its inverse-designed approach also offers scalable and full compatibility with single-laser continuous-wave operation at 1550 nm. Experimental results show that MDTransformer achieves 40.4% area reduction, 63.6% power saving, 40.6% energy saving, and comparable latency over the state-of-the-art PTA across different workloads (i.e., DeiT-Tiny/Small/Base and BERT-Base/Large). These results show that MDTransformer offers a practical solution for high-performance and energy-efficient transformer-based systems.

Self-play in simulation produces robust driving policies at scale. Demonstrations of such behavior have been made using privileged vectorized observations such as exact poses and velocities, even for occluded agents. This assumes that perception is solved and introduces a representation gap with the partial observation of a deployed agent driving from the perspective view of egocentric cameras. A common fix, distilling the privileged policy into a camera-input student, leaves the student imitating decisions its own view cannot justify. Instead, we establish perspective-view self-play as a practical training regime. We introduce Pictura, a GPU-accelerated multi-agent driving simulator that renders each agent's egocentric view at every step, mitigating the representation gap at its source. Pictura sustains up to 500K agent-steps/s (2M images/s) on a single H100. Using Pictura, we train Alberti by self-play with plain PPO. It is the first large-scale driving self-play policy trained directly from perspective images, without privileged observations. Training spans 50B agent steps for ~35M km of driving. It approaches the driving performance of its privileged vectorized counterpart, and transfers zero-shot to Waymo Open Motion Dataset layouts re-rendered in Pictura, where it outperforms privileged vectorized agents. Project page: https://valeoai.github.io/Pictura/

Tabular Foundation Models (TFMs) have emerged as novel approaches for tabular predictive tasks, demonstrating competitive predictive performance to ensemble tree-based models. Most TFMs are trained and evaluated on independent and identically distributed data, but this assumption changes in real-world scenarios due to distribution shifts, which compromise the robustness of models. Limited research has been conducted of TFMs under distribution shifts. We present an empirical evaluation of Out-Of-Distribution (OOD) performance of nine TFMs, spanning diverse pre-training strategies and architectures: TabPFNv2, TabPFNv2.5, TabPFNv2.6, TabPFNv3, TabICL, TabICLv2, Mitra, LimiX and TabFM. Three real-world datasets from the TableShift study were considered (HELOC, Voting, Childhood Lead), covering label, socioeconomic, and geographic shift types. Our results show that all evaluated TFMs degrade systematically under distribution shift regardless of pre-training strategy, with shift gaps ranging from 0.003 to 0.060 depending on shift type. The relationship between in-distribution and OOD predictive performance documented for classical tabular models extends into TFMs. We also identified a scalability gap, as high-performing models demand significant memory and computational resources beyond what standard deployment infrastructure can support. This study extends existing benchmarks for OOD in tabular data, providing evidence to support their adoption in high-stakes domains characterized by structural distribution shifts.

Ultra-high-resolution (UHR) remote-sensing (RS) imagery provides fine-grained Earth-observation evidence over city-scale scenes, but poses a fundamental challenge for multimodal large language models (MLLMs): task-relevant evidence is often sparse, local, and spatially dispersed across extremely large visual contexts. A natural solution is to equip MLLMs with zoom-in tools for active local inspection. However, through a pilot study on XLRS-Bench, we find that zoom-in is only partially effective: it resolves easy and medium-level tasks with locally recoverable evidence, but saturates on hard cases requiring global search, multi-region comparison, path planning, or dispersed-evidence reasoning. Motivated by this finding, we move beyond single-tool zoom-in and introduce GeoMTVR, a large-scale Geospatial Multi-Tool Visual Reasoning dataset built from wide-area satellite imagery. GeoMTVR contains 13K UHR VQA samples with interleaved reasoning trajectories, diverse visual tool calls, and returned visual observations, enabling models to learn question decomposition, tool selection, regional inspection, object-level grounding, auxiliary visual reasoning, and cross-tool evidence integration. Beyond supervised fine-tuning, we propose a tool-attention-focused reinforcement learning algorithm that concentrates optimization on critical tool-use decisions, including when to invoke tools, which tool to select, where to apply it, and how to interpret tool outputs. By combining SFT on GeoMTVR with our RL algorithm, we develop GeoLens, a multi-tool visual reasoning MLLM for UHR RS. Experiments show that GeoLens consistently outperforms direct reasoning and single-tool zoom-in baselines, achieving stronger accuracy, better evidence grounding, and more efficient tool-use trajectories.

Recently, memory management has become a key infrastructure for LLM-based agents, as it directly affects long-horizon reasoning, personalized responses, and knowledge reuse. However, existing LLM memory systems typically adopt a coarse-grained (utility-agnostic) manner that treats heterogeneous user-LLM interaction records uniformly, leading to redundant and low-impact records persisting in the memory repository. To address this challenge, we present MemLens, a value-aware memory management system that takes memory records as first-class data objects. MemLens provides an end-to-end interactive analytics dashboard that exposes the complete memory lifecycle, including Shapley-style memory evaluation, value-aware storage, and memory-assisted response. Through a study-copilot application, the system enables users to inspect memory values, visualize hierarchical memory structures, and compare various memory management strategies in terms of response quality, retrieval latency, and token consumption. Therefore, our MemLens can serve as an efficient, interpretable, and personalized long-term memory management system for LLM-based agents.

The vision of self-driving networks that monitor, reason, and act upon themselves with minimal human intervention relies on tightly coupled monitoring, analytics, and actuation functions. In this work, we treat these functions as three operational macro-intents: continuous telemetry, real-time analytics, and programmatic actuation, and formalize the health of each function as an intent that the network must continuously satisfy. A critical, yet underexplored, challenge stems from the causal coupling among these intents, where a singular fault within one macro-intent propagates as a co-drift and subsequently triggers cascading, symptomatic anomalies across the remaining intents. This ambiguity makes it exceedingly difficult for existing, reactive approaches to distinguish the true root-cause intent from symptomatic victim intents, and their reliance on threshold-crossing detection leaves insufficient time for proactive remediation. We introduce MILD, a novel framework that reformulates intent assurance from reactive drift detection to proactive failure prediction. Grounded in our three-macro-intent formulation of the self-driving control loop, MILD employs a teacher-augmented Mixture-of-Experts architecture with a hybrid objective that jointly optimizes intent failure prediction and root-cause attribution. MILD enables KPI-level diagnostics via SHAP explainability and dynamic intent failure urgency estimation via multi-horizon modeling. Our extensive evaluation of MILD across three environments of increasing realism, from a controlled statistical benchmark, to a microservices application, to an SDN-based edge-to-cloud testbed, demonstrates that MILD achieves high failure detection rates, strong remediation lead times, and accurate intent-level root-cause disambiguation. This positions MILD as a practical enabler of closed-loop assurance in next-generation autonomous networks.

Unmanned aerial vehicles (UAVs), particularly quadcopters, present unique challenges for autonomous control due to their underactuated dynamics: only four available control inputs must govern six degrees of freedom. This paper investigates a physics-aware, end-to-end deep reinforcement learning (DRL) approach that acts directly on low-level body inputs, total thrust and body torques $(T, τ_x, τ_y, τ_z)$, and closes the loop through a high-fidelity Simulink environment. Our simulator integrates a 12-state rigid-body model (MATLAB Level-2 S-Function) with (i) an Action2RPM allocation based on the Moore-Penrose pseudo-inverse of a coefficient matrix derived from thrust and drag terms, and (ii) first-order actuator dynamics for each motor (time constant $T_m = 0.076$ s), including rotor gyroscopic coupling. A shaped reward balances goal-reaching and stability using an exponential position well, attitude penalties, and quadratic velocity costs. Four DRL algorithms, DDPG, TD3, PPO, and SAC, are evaluated in two stages: (S1) thrust-only hover and (S2) hover with pitch torque and a translated goal. Results show that SAC and TD3 achieve superior stability and exploration efficiency, while PPO is less sample-efficient. The study highlights the significance of modeling actuator lags and aerodynamic moments for stable low-level control and provides a reproducible benchmark for quadcopter DRL.

Recent generative models can produce images with few obvious visual artifacts, weakening detectors and explanations that rely only on surface appearance. We present LaP-Forensics, a multimodal framework that augments RGB semantics with reconstruction-based forensic evidence. A frozen Stable Diffusion DDIM inversion-reconstruction model provides a fixed reconstruction reference, and its residual map measures local compatibility with that reference. Independent projectors encode the RGB image and residual map before a structured Where-What-Why model predicts a textual analysis and an artifact mask.Supervised fine-tuning is followed by Group Relative Policy Optimization (GRPO), whose reward combines mask overlap with output-structure and evidence-reference terms. These text-side terms encourage the model to refer to the consistency map but do not constitute a verifier of free-form textual truth. A separate image-level head fuses RGB and DDIM-residual class features. Experiments show cross-generator detection on UniversalFakeDetect and competitive artifact localization on the official SynthScars benchmark. Controlled cue-construction, inversion-horizon, component, reward-term, and counterfactual analyses support the utility of the residual stream under the evaluated settings, while free-form textual faithfulness and reliability under post-processing remain open limitations.

Wikipedia and Wikidata are widely used for information access, LLM pre-training, and retrieval-augmented generation. Their knowledge is deeply connected but scattered across text, tables, and knowledge graphs. This raises a practical question: when these modalities disagree, how can we detect and explain the conflict? We study this problem as \emph{modality-level inconsistency detection}. We first introduce a taxonomy of cross-modal knowledge inconsistencies, covering information granularity differences, direct conflicts, temporal changes, and KG incompleteness. We then present \textsc{Kontrast}, an automatic framework that uses Text-to-SPARQL and LLM reasoning to compare table-based answers with KG evidence and categorize the resulting inconsistencies. Experiments on various Table-QA datasets show that cross-modal inconsistencies are common and informative. They reveal not only true knowledge conflicts, but also missing KG structure and temporal mismatches while being limited by Text-to-SPARQL errors and noise. Our analysis shows that text, tables, and KGs can complement and correct one another through systematic comparison. \textsc{Kontrast} provides a practical tool for large-scale knowledge auditing and establishes a benchmark for future work on cross-modal knowledge consistency. Code and data are available at https://github.com/ECLADATTA/KONTRAST.

As LLMs increasingly mediate the political information citizens rely on, there is still no standardized way to assess whether they do so responsibly. We introduce Polistemics, a theory-grounded benchmark for evaluating LLMs as mediators of political information in elections. Prior work has treated this task as reproduction rather than mediation, leaving its epistemic dimensions and interaction with imperfect information unaddressed. We ground the evaluation in Epistemic Modesty, a normative standard derived from citizens' epistemic agency, and test it across controlled settings that vary informational properties such as clarity, noise, and consistency. Applying the benchmark to three state-of-the-art LLMs on the 2025 German and Dutch elections, we find that high aggregate scores mask systematic failures. Models mediate reliably under clear evidence but break down under absent, vague, or contradictory information, while flattening the intensity of political language. These failures are likely driven by party priors, influenced by party labels and output language. Reliable mediation appears achievable, but no model delivers it consistently.

Any-to-any models predict any modality from any combination of others within a single network, a formulation used in multimodal vision and vision-language models, and increasingly in scientific domains such as ecology and astronomy. Existing any-to-any models are typically trained from scratch using encoder-decoder or diffusion architectures, impacting their performance and preventing them from using strong pre-trained decoder-only models as a prior. In this work, we investigate decoder-only any-to-any multimodal modeling, which treats all modalities symmetrically and supports arbitrary modalities as inputs and outputs without modality-specific heads, losses, or task pipelines. Because every modality is both an input and an output of the same model, the resulting model, named Modus, can support a range of applications, such as chained generation through intermediate modalities or cross-modal self-verification by scoring the model's own outputs with another generated modality. Modus demonstrates strong out-of-the-box performance and is competitive with specialist and multitask baselines using a single model across various benchmarks. All materials are open-sourced at https://modus-multimodal.epfl.ch/.

Question answering (QA) over irregular clinical time series (ICTS) plays a pivotal role in a wide range of healthcare applications. Although recent multimodal time-series large language models (LLMs) have shown considerable promise in general-purpose time-series QA, they remain poorly equipped to model the sparsity, asynchrony, and irregular sampling patterns of clinical observations. To fill this gap, we propose ClinPRISM, a cost-effective multimodal LLM reasoning framework for question answering over ICTS data. First, we devise an irregularity-aware multi-scale encoder to capture sparse clinical evidence at diverse temporal scales. Then, we propose a temporal evidence distiller to integrate representations across these scales and compress them into a small number of LLM-compatible tokens. Moreover, we introduce a progressive alignment strategy that sequentially aligns the irregular trajectories with the LLM's textual embedding space. To facilitate training, we construct 30,000 clinical time series paired with multi-scale descriptions, together with 41,000 instruction-tuning instances spanning 11 tasks. Using a 4-billion-parameter LLM backbone, ClinPRISM achieves state-of-the-art performance on the held-out evaluation benchmark while using only 16 time-series tokens and achieving an average inference latency of 0.15 seconds per question.

Clinical diagnostic evaluation should not only assess whether models can provide correct diagnoses, but also reflect the realities of clinical practice, including progressive disclosure of multimodal information, dynamic updating of diagnostic hypotheses, and continuous refinement of clinical reasoning. However, existing evaluations of multimodal large language models (MLLMs) typically rely on single-turn or isolated tasks, making it difficult to fully capture the complexity of real-world clinical diagnosis. To bridge this gap, we developed ClinMM-Bench, the largest multi-turn multimodal clinical diagnostic evaluation benchmark to date. ClinMM-Bench contains 1,089 challenging real-world clinical cases and 3,760 medical images across eight specialties. We systematically evaluated 15 representative MLLMs using a two-level evaluation framework that assessed both diagnostic accuracy and diagnostic reasoning quality. Results showed that proprietary models achieved the highest overall diagnostic accuracy, but the proportion of completely correct diagnoses remained limited across all models. In terms of diagnostic reasoning quality, current models can identify plausible diagnostic directions but still have considerable limitations in generating reliable diagnostic reasoning. Error analysis further identified five representative failure modes: information synthesis failure, knowledge mapping error, perception error, premature closure, and visual hallucination.

Face de-identification (De-ID) aims to remove or conceal personally identifiable facial features in images or videos to prevent identity recognition while preserving utility for downstream tasks. With the rising emphasis on data privacy and responsible AI, face De-ID has emerged as an active research area spanning computer vision and privacy-preserving communities. Early approaches, and many contemporary ones, operate in the digital domain by modifying pixel-level or appearance-level features through post-capture processing. Recent advances extend face De-ID beyond post-processing by integrating privacy mechanisms directly into sensors during image acquisition, bridging sensing systems and downstream vision algorithms. In parallel, physical-domain methods explore wearable accessories and materials that conceal identity information in real-world environments prior to capture. In this survey, we present the first unified overview that spans the full data acquisition pipeline, encompassing the physical, sensor, and digital domains. Through this domain-centric lens, we systematically analyze current methodologies, technical progress, and the distinct challenges inherent to each stage. We then review and organize existing evaluation protocols, examining current practices and highlighting the critical need for standardized, comprehensive benchmarks. Finally, we identify key open problems and outline emerging research directions to guide future work in this rapidly evolving field. To support ongoing research, we maintain a project page that organizes relevant literature with collected datasets and open source code: https://github.com/CV-AC/Awesome-FaceDe-ID.

In this work, we study the use of Vision-Language Models (VLMs) for anomaly detection in an agent-driven game Quality Assurance (QA) pipeline focusing on geometry clipping. In this evaluation, a custom exploration agent navigates a game level to collect visual observations, while the automatic annotation pipeline provides frame-level clipping labels. This setup allows us to evaluate recent VLMs on a controlled anomaly detection task without manual annotation. We benchmark six recent VLMs (Gemini, GPT, Qwen, Gemma, Llama, and Ministral) under a zero-shot prompting setting and analyse their sensitivity to four prompt variants. Our results show that while the VLMs can capture visual cues associated with geometry clipping, they all produce substantial false positives on visually ambiguous frames such as near-contact geometry and partial occlusions. Gemini-3.1-Flash achieves the best overall accuracy and is the most robust to prompt variation, while open-source models exhibit large precision--recall swings depending on the prompt design. These findings suggest that current VLMs are best suited as high-recall candidate filters within multi-stage QA pipelines rather than as standalone bug detectors.

World-Action Models (WAMs) augment robot policies with future visual prediction, but it remains unclear what the visual modality should learn for control. While photorealistic future prediction provides dense supervision, it also incurs substantial computation and can allocate capacity to texture, illumination, and background variations that are only weakly related to action selection. Recent efficient WAM variants suggest that the main benefit of the video branch may not lie in the rendered future itself, but in the control-relevant visual representations induced during training. In this work, we revisit future video prediction from a dynamic-centric perspective and ask whether an existing RGB-based WAM can be redirected from appearance-dominated reconstruction toward interaction-induced visual dynamics without introducing additional modality-specific predictions or online inputs at deployment. We propose DC-WAM, a dynamic-centric WAM framework that redistributes supervision and computation in the RGB video branch. At the supervision level, DC-WAM combines temporal-difference flow matching with trajectory-guided weighting, emphasizing dense temporal changes and localized regions where the gripper, manipulated objects, and contact areas move. At the reasoning level, DynaRoute predicts token-wise dynamic relevance and converts it into an attention bias, guiding the model toward control-relevant future tokens. Experiments in simulation and on real-world manipulation tasks show that DC-WAM consistently improves policy performance, especially under out-of-distribution perturbations in lighting, object appearance, and background texture.

Complex structured reasoning tasks often require additional computation, yet current language models obtain it mainly by increasing parameter scale or by serializing intermediate steps as chain-of-thought (CoT) tokens. The former raises training and deployment costs, while the latter ties reasoning computation to autoregressive output length. We introduce Penelope, an efficient latent-reasoning framework for pretrained decoder-only Transformers that localizes recurrent computation to a selected decoder interval. The lower decoder prefix is evaluated once to construct a problem-conditioned boundary memory, which is then iteratively refined through time-modulated GRU dynamics and recurrent readout states before answer generation. A progressive CoT-to-latent curriculum transfers visible reasoning into this internal recurrent path, allowing additional computation to be allocated in latent space without repeatedly executing the complete decoder or generating a long intermediate trace. Experiments on open-source structured-reasoning benchmarks show that, at validation-selected latent budgets, Penelope attains competitive accuracy relative to established latent-reasoning models while reducing measured inference latency. These results show that latent refinement can be localized to a narrow decoder interval, reducing repeated full-decoder execution without generating a long visible reasoning trace and providing a practical accuracy-efficiency tradeoff for decoder-only Transformer models.

Autonomous Network Levels 4-5 require AI agents to invoke tools across vendor boundaries without human oversight, yet existing management standards lack a standardized mechanism for cross-vendor trust visibility. When a tool from Vendor B is compromised, agents from Vendor A continue invoking it -- unaware of the trust degradation -- causing cascading service impact. We present AgentToolMO, a proposed 3GPP NRM information model for agent tool trust management. The model comprises: a formally defined trust state machine with provable graduated enforcement, damped cascade propagation with bounded convergence, cross-vendor trust notifications via existing Management Services (MnS) interfaces, and retroactive impact assessment through NRM dependency graph traversal. Simulation-based evaluation across multi-vendor topologies shows that standardized cross-vendor notifications reduce blast radius from hours-scale undetected propagation to near-real-time containment bounded by MnS notification delivery, with cascade convergence guaranteed in bounded iterations and sub-linear notification scaling across vendor domains. The framework operates within existing 3GPP management infrastructure, leverages existing protocols, and provides a standardization pathway for trustworthy multi-vendor autonomous network management.

Vision-Language-Action (VLA) models have shown strong potential for general robot manipulation, but most existing models rely on 2D visual-language backbones and lack fine-grained 3D understanding of target objects, especially under occlusion, pose variation, scale changes, and precise spatial interaction. We propose an object-centric 3D representation alignment framework built upon $π_0$, using SAM3D as a frozen 3D teacher to provide target-object 3D priors during training. Specifically, we localize task-relevant objects with object recognition models, generate corresponding object masks, and use SAM3D to extract dense object-level 3D representations, which are aligned with intermediate visual features of $π_0$. This enables the policy to internalize target-object 3D information while preserving the original RGB-language-to-action inference pipeline without requiring depth, point clouds, masks, SAM3D, or additional 3D modules at test time. Simulation experiments show consistent improvements, achieving 99.1\% on LIBERO and an average length of 4.11 on CALVIN. Real-world experiments further demonstrate that our method is particularly effective in long-horizon manipulation scenarios where the robot must focus on different target objects across multiple subtasks.

Annotation is among the most demanding visualization tasks to automate, as it simultaneously requires correctly navigating visual, semantic, and stylistic constraints. Failure to meet any of these conditions severely undermines the utility of an annotation, rendering it challenging to read, inaccurate, or visually discordant. Despite a growing body of annotation tools and automations, no existing benchmark or evaluation framework tests whether these conditions are met because of their scope and annotation not being the focus of their studies. We introduce AnnoBench, a benchmark for visualization annotation that materializes the inherent challenges of this domain in a structured and testable manner. AnnoBench pairs visualizations from professional data journalism and visualization galleries with annotation tasks, spanning four representation formats, five chart description conditions, and two prompt specification levels. The benchmark is executed via VLM-as-a-judge, using models aligned with manual human assessment. We evaluate the benchmark via four one-factor-at-a-time experiments, exploring the effects of input representation, semantic context, and prompt specificity, and model selection on annotation quality. This work provides a foundation for advancing annotation automation, tooling, and visualization-generation pipelines.

Activation steering controls model behavior by editing internal activations at inference time. We study its input-side dual: optimizing a fluent prompt so that a chosen internal latent is driven toward zero, with no inference-time model access. Our target is an "evaluation-awareness" latent-linearly readable and steerable in recent work-whose control would threaten the validity of safety evaluations if models behave differently when they detect being tested. Adapting Fluent Dreaming / EPO with a negated feature term (GCG-style token optimization plus a self-cross-entropy fluency regularizer, swept over a fluency weight), we suppress the latent under five target constructions-a CAA direction, a subspace norm, an SAE feature, a single MLP neuron, and a behavioral logit-on Llama-3.2-3B and Llama-3.1-8B. The latent is robustly suppressible ($z\approx-7$), and a causally-validated Llama Scope SAE feature can be fully and selectively turned off. But our controls tell a cautionary story about the CAA direction: a placebo random direction is suppressed just as hard and shifts behavior just as far, and when we hold a real eval passage in context and optimize only a prefix, suppressing the eval-direction fails to reduce-and slightly increases-the model's behavioral eval judgment. Activation-readability, in short, is not behavioral controllability. We further find that a single MLP neuron is eval-correlated but not causal at both scales, and that scanning the real Pile yields a natural-text baseline competitive with the optimizer for the internal direction. A positive control validates our erasure detector, bounding an erasure-vs-rotation question earlier left open.

Graphical user interface task evaluation aims to determine whether a GUI agent has successfully completed a user instruction. Automated GUI task evaluation has received increasing attention because the evaluation results can serve as reward signals for both test-time scaling and post-training. However, reliable GUI task evaluation remains challenging because the judgments often require access to environment states, such as system configurations, file data, and application settings, beyond the screenshots of execution trajectories. In this paper, we propose an interactive reward agent (IRA) based on a propose-then-verify framework to acquire and verify evidence from the post-execution environment. Given a task instruction and a GUI environment after the GUI agent execution, IRA first proposes the task completion conditions and then verifies them by invoking system tools, application tools, and GUI tools. This design combines evidence from both visible interfaces and the environment state in an interactive process. We further introduce GUI-RewardBench, a benchmark of 321 GUI task trajectories spanning 10 Ubuntu desktop application categories. Experiments show that IRA achieves 86.9% accuracy on GUI-RewardBench, outperforming existing evaluator baselines. We further apply IRA to reinforcement learning of GUI agents, achieving a 34.0% OSWorld success rate, which demonstrates that IRA can provide effective reward signals for training GUI agents.

Learning deployable manipulation policies is bottlenecked by the scarcity of data that is both high-fidelity and scalable. Real-robot teleoperation is accurate but costly to scale; robot-free UMI capture scales readily, and current practice uses the resulting data mainly for pre-training, adding a small real-robot "anchor" at post-training. We ask whether raising the fidelity of robot-free UMI data, rather than shrinking the real-robot fraction, can remove that anchor. We present HiFi-UMI, a portable UMI data-production system co-designed for trajectory accuracy, inter-gripper relative pose, synchronization, and field of view: head-mounted offline stereo-inertial SLAM, native rather than reconstructed relative pose, a shared microsecond GPIO trigger, and two wide-angle cameras per hand covering ~200 degrees. It reaches 3 mm workspace-local end-effector accuracy without external tracking infrastructure. Using this corpus, we demonstrate zero-robot post-training: a policy post-trained solely on HiFi-UMI demonstrations deploys directly on a real robot and matches in-domain teleoperation across three backbones spanning the vision-language-action and world-action-model families, with success-rate differences of -2.5, +3.1, and -0.6 percentage points on StarVLA-QwenPI, OpenPI-pi_0.5, and LingBot-VA; the strongest policy reaches 85% on a precision insertion task, even though the teleoperation baseline is collected in the evaluation scene and no HiFi-UMI trajectory is. Pre-training on 4,000 hours from the same corpus lowers action error on ten unseen tasks by 41% and, on StarVLA-QwenPI, raises real-robot success by a further 18.1 percentage points. We open-source HiFi-UMI-2K, 2,000 hours of microsecond-synchronized, ultra-wide-FoV demonstrations, each automatically reconstructed and validated through simulation replay, as a large-scale, high-fidelity resource for the robot-learning community.

Evaluating AI agents in interactive environments is hindered by fragmented tasks, scaffolds, verifiers, and scoring rules. Existing efforts focus on narrow settings, remain limited in scale, or require costly reruns, leaving much of the empirical record incomparable. We introduce Messier, a unified corpus of 957,253 records that span 30 benchmarks, 714 agents, 11,891 tasks, and 74,205 verifiers. Messier consolidates public benchmark scores and supplements them with five-agent runs across six underrepresented professional and scientific domains, including a recent legal benchmark. Each record is standardized by model, scaffold, environment, task, verifier, and aggregation rule, with SOC/NAICS classifications for occupational and industry analysis. Using this corpus, we show frontier progress is uneven across benchmark types, with "function calling" saturated, "programming" improving the fastest, and "enterprise workflows" remaining the most challenging. Furthermore, counterfactual rescoring shows that strict all-pass aggregation in multi-verifier tasks can obscure progress and artificially alter agent rankings. From these standardized records, we derive capability scales that align with Epoch's Evaluation Capability Index rankings at Spearman \r{ho} = 0.81 and can be specialized by domain, occupation, action space, or verifier type. Messier provides a foundational, reusable infrastructure for agent capability scaling, benchmark auditing, and fine-grained analysis of evaluation failures.

AI coding agents are being adopted at historic speed, yet security and risk concerns remain the primary barrier to scaling agentic AI across organizations. Existing security controls for coding agents are not systematically distributed to engineering teams, and vendor-native solutions introduce ecosystem dependencies that may not suit every deployment context. This paper investigates whether off-the-shelf security controls can be implemented on commercial AI coding agents and scaled to a distributed user base via a custom agent harness. A phased testing methodology was applied across four agent configurations --- two commercial agents with and without controls, a baseline harness, and a security-hardened harness --- using a 23-test suite derived from the OWASP Top 10 for Agentic Applications. SHarD (Secure Harness Distribution), a distributable harness built on the Pi agent harness, demonstrated that three categories of security controls --- OS sandboxing, skill scanning, and tool restriction --- can be embedded and distributed via a single install command while retaining equivalent efficacy to direct installation on commercial agents. SHarD achieved an adjusted score of 100\%, matching the best securely configured commercial agent, with no regression across any test category. Notable observations include evidence that model non-determinism produces inconsistent security outcomes and that autonomous agent behavior can cross system boundaries in ways that OS sandboxing directly mitigates. Initial characteristics toward a control harness fitness framework are proposed, and a third research question is identified for future investigation.

Recursive self-improvement requires turning evidence of model failures into better models. Data-centric post-training research entails diagnosing capability gaps, designing and validating training-data strategies, and learning from checkpoint feedback. Can LLM agents automate this loop? Existing benchmarks entangle research decisions with optimization, serving, evaluation, and systems implementation, obscuring agents' research capability. We introduce RSIBench-Data, a controlled benchmark of LLM agents as data-centric researchers with a fixed post-training stack. Agents iteratively revise training-data strategies for a fixed target model; training and serving use Tinker-backed services, official evaluation runs through Harbor and E2B sandboxes, and budgets are fixed across agents. We evaluate four frontier agents on six benchmarks across software engineering, terminal use, scientific question answering, and mathematics. Agents demonstrate core data-centric research capabilities: in 58.33\% of settings, they improve upon the first valid attempt by refining strategies from feedback. However, improvement is inconsistent. Among searches continuing after the best observed score, 78.26\% end with a lower-scoring final attempt, while the rest only recover the same peak. A strong candidate may therefore appear early or midway through a run even as later revisions fail. Trajectory analysis identifies four patterns in stronger runs: accurate hypotheses, validation-grounded supervision, behavior-aligned data, and preservation of strong checkpoints. These findings suggest that current agents can make useful data-centric discoveries but cannot yet translate feedback into consistent improvements. RSIBench-Data provides a measurable, auditable testbed for the research capabilities required for recursive self-improvement. We open-source our code at https://github.com/evolvent-ai/RSIBench-Data.

This paper investigates how multi-agent systems (MAS)-based on large language models (LLMs) can support actuarial risk modelling, with a particular focus on uncertainty quantification. Actuarial workflows represent a high-stakes decision-support setting where unreliable outputs may lead to incorrect risk assessment, unfair pricing, and regulatory non-compliance. To address uncertainty introduced by the probabilistic nature of LLMs and dependencies between agents, a multi-agent framework is proposed in which specialised agents perform data preparation, modelling, review, and explanation tasks under a central hub. The main contribution is a novel approach to uncertainty propagation using token-level log-probabilities and a Bayesian Network. Importantly, log probabilities are not treated as direct probabilities of correctness or task success. Instead, length-normalised log-probability summaries are transformed into calibrated task-level confidence estimates before incorporation into the Bayesian Network. Results show that the framework reproduces baseline actuarial performance while providing additional insight into workflow stability and runtime uncertainty propagation.

Quantum error correction (QEC) is indispensable for scalable fault-tolerant quantum computing. However, discovering QEC codes that remain effective is challenging, as logical performance depends on the interplay between code structure, hardware, syndrome extraction, and decoding, which often impose competing requirements. Here we introduce OmniQEC, an efficient AI scientist for discovering QEC codes suited to deployment on modern quantum processors. OmniQEC formulates QEC design as an iterative discovery process in which an orchestrator, implemented by advanced large language models (LLMs), coordinates code generation, code-level screening, syndrome-extraction synthesis, and decoder-based circuit evaluation. At its core, OmniQEC combines a self-evolving reasoning mechanism with a slow--fast synergistic workflow: a fast loop explores candidates using inexpensive code-level proxies, whereas a slow loop performs physically grounded circuit-level evaluation and feeds the resulting evidence back into the search. We evaluate OmniQEC across four qLDPC construction families, three LLM backends, and $14$ total-physical-qubit budgets per backend. The discovered codes show steadily improving logical-error suppression with increasing physical-qubit budgets and outperform the BB codes with $[\![72,12,6]\!]$ and $[\![144,12,12]\!]$ under complete-implementation budgets of 98 and 240 physical qubits, respectively. The discovered codes are hardware-friendly and may be of independent interest for practical QEC implementation. These findings pave the way towards LLM-assisted QEC discovery grounded in physically informed code--circuit--decoder co-design.

Many time-series forecasts depend not only on prior observations but also on actions specified during the forecast period. In intensive care units (ICUs), future vital signs and laboratory values are influenced by treatments such as vasopressors. However, models that predict the full future sequence all at once make little use of these treatments, whereas autoregressive models can accumulate errors. We introduce DRIFT, a hybrid framework in which a direct model produces the primary forecast and a recursive, action-conditioned model contributes constrained corrections. We evaluate DRIFT on 6,046 admissions from MIMIC-IV and 8,345 admissions from eICU-CRD. Averaged across the 8-, 24-, and 48-hour forecast endpoints, DRIFT reduces mean absolute error for mean arterial pressure (MAP) by 0.673% relative to an action-conditioned Temporal Fusion Transformer (TFT-action) on MIMIC-IV and achieves the lowest corresponding error among the compared models on eICU-CRD. Although the overall accuracy improvement is modest, a MIMIC-IV audit restricted to windows in which the supplied treatment sequence was altered showed that DRIFT achieved lower observed-target MAP error than TFT-action at 8 and 24 hours. Treatment-sequence alteration increased DRIFT's MAP error by 0.21-0.26 mmHg more than it increased TFT-action's error, with prediction changes occurring primarily after the supplied paths diverged. In a separate robustness experiment, the MAP advantage persisted under three shared checkpoint-selection rules emphasizing overall endpoint error, MAP error, or both equally.

We introduce Shieldstral, a 3B-parameter policy-adaptive multimodal safety classifier that matches or outperforms models nearly 7$\times$ its size on text safety benchmarks and sets a new state of the art on multimodal safety classification. Shieldstral formulates content moderation as a binary question-answering task. This simple formulation unifies diverse moderation tasks into a single yes/no problem, enabling heterogeneous safety datasets with divergent taxonomies to be consolidated under one training framework. We present the data construction recipe, covering curation and generation of approximately 54.1M samples and a fine-grained evaluation set to evaluate policy adaptability. Together, these enable a small adaptive model to match or outperform much larger models.

Skills have become an important abstraction for enabling large language model (LLM) agents to reuse past experience in long-horizon interactive tasks. However, existing trajectory-to-skill methods often produce flat collections of high-level textual skills that are stored and retrieved independently, leaving skill relations underutilized and maintaining a gap between high-level skills and executable actions. In this paper, we propose HiSkill, a hierarchical skill graph framework that organizes interaction trajectories into a directed graph with skill nodes, AtomicOp nodes, and typed edges. Specifically, the graph connects reusable high-level skills with executable action templates, while also capturing decomposition, temporal transition, compatibility, support, and recovery relations among them. At inference time, HiSkill retrieves a compact task-relevant subgraph and performs subgraph-guided task execution, where a symbolic task state, an active skill, and the retrieved subgraph guide the LLM agent to switch skills, select AtomicOps, and ground executable actions iteratively. Experiments on three interactive environments show that HiSkill outperforms state-of-the-art baselines while reducing inference token consumption, demonstrating the effectiveness of bridging high-level skills and executable action grounding through a hierarchical skill graph. Our data and code is available at https://github.com/BUPT-GAMMA/HiSkill.

This paper describes the participation of team "Go To Germany" in the ImageCLEF 2026 Deepfake Detection and Generation Task. For the image generation task, we employ FLUX.1-dev with PuLID for identity-preserving face synthesis, combined with a multi-model PGD adversarial attack targeting 12 detectors simultaneously (DiffJPEG-in-loop, MI/DI/EoT, adaptive weighting, two-stage warm-start). Our approach achieved 90% evasion against organizer detectors and 57.6% against participant detectors, with a final generation score of 0.4170. For the image detection task, we combine two complementary detectors - SigLIP+DINOv2 for AI-generated images and GenD-DINOv3 for face manipulations - in a max-probability ensemble, achieving 99.4% accuracy on baseline deepfakes but suffering from high false-positive rates on real images, resulting in a final detection score of 0.6986. Beyond the official submission, we conducted a self-initiated investigation of purification-based adversarial detection, comparing three families of detection signals across six detectors that share a CLIP ViT-L/14 backbone. We find that raw $|Δ\text{logit}|$ under median-3 purification, applied through the EFFORT detector, separates adversarial inputs from clean inputs with AUROC 0.81-0.98 across four adversarial source types - a finding that refutes the simple backbone-preservation hypothesis and exposes a sharp JPEG-quality cliff at Q70 where the signal collapses.

Distributed constraint optimization problems (DCOPs) provide a popular framework for distributed decision making under limited communication, but many real-world instances are too large to solve monolithically. We address this challenge from two complementary directions. We revisit the connection between DCOPs and potential games, and adapt modern online learning algorithms for equilibrium finding to DCOPs. We show that these algorithms are competitive with representative incomplete DCOP algorithms. We then turn to decomposition frameworks for large-scale DCOPs, motivated by large-scale decentralized satellite scheduling. We propose a new framework that separates a DCOP into two interacting subproblems: a high-level meta-DCOP for task allocation, and independent local optimization problems for scheduling. To couple the two levels, we develop a novel iterative pricing method that updates the meta-level utilities using feedback from the local optimizers. Combining our online learning methods with our iterative pricing framework, we obtain near-optimal performance on real-world decentralized satellite scheduling problem instances, fulfilling over 99% of observation requests compared with 87% for state-of-the-art baselines.

Quantum computers are moving from research laboratories to industrial machines accessible via the cloud and integrated into high-performance computing facilities. However, translating theoretical quantum protocols into hardware experiments remains a major bottleneck, requiring expertise across protocol design, compilation, simulation, and cloud execution. Here, we introduce an agentic workflow that automates this pipeline on neutral-atom quantum processors (here two Pasqal QPUs available on the cloud) while keeping the researcher in the loop for critical validation. In three case studies from many-body physics and optimization, the agent went from published paper or patent to a QPU campaign run overnight. In particular, human intervention was crucial to ensure scientific validity: the agent selected an inadequate observable in one experiment and constructed a plausible but incorrect hardware diagnosis in another, with both failures detected only through domain-expert review. Finally, we use a second agent to classify a corpus of 633 Rydberg-array arXiv papers and show that nearly half are implementable on present-day QPUs while identifying specific hardware upgrades needed for the rest. Together, these results demonstrate that agentic workflows provide a practical bridge between theoretical ideas and physical hardware, opening quantum experimentation to a much broader scientific community.

Deep learning models in computer vision face significant challenges when trained on long-tailed datasets, where a few majority classes dominate while many minority classes are severely underrepresented. Such imbalances frequently arise in real-world scenarios such as rare species recognition, manufacturing fault detection, and medical image understanding, leading to biased models that underperform on tail classes. Existing reweighting methods typically rely on static class frequencies to penalize the model, ignoring the dynamic nature of how effectively a network actually learns a class over time. We address this by introducing a novel Learning-Dynamics Aware Loss (LDAL) function that shifts the focus from static sample counts to dynamic learning progress. LDAL framework adjusts class weights continuously by leveraging: (i) the strength of learned feature representations (semantic scale), (ii) the intrinsic learning difficulty of each class, measured via the Shannon entropy of its predictions, and (iii) an inter-epoch regularizer term that tracks prediction shifts between consecutive epochs to stabilize training and avoid local minima. LDAL is purely a objective function which incurs negligible computational overhead while adapting to the feature learning of the model. Experimental results on multiple benchmark datasets demonstrate that our approach significantly surpasses state-of-the-art reweighting loss functions, providing an optimal trade-off between accuracy and generalizability. The source code is available at https://github.com/sdm2026/ldal

Food image segmentation plays a vital role in health-related applications such as nutrition tracking and personalized health monitoring. However, existing models often underperform on visually similar ingredients and rare food categories. To address this issue, we propose two plug-and-play multimodal modules that enhance the segmentation performance by leveraging ingredient labels inferred from food images using large language models (LLMs). The first module, called LIM-F (Language Injection Module for Features), is designed to pair with any image encoder that produces multi-layer outputs (e.g., Swin Transformer), while the second module, LIM-Q (Language Injection Module for Queries), targets Mask2Former-style Transformer-based decoders. Both modules enable training without the need for pre-aligning images with text by directly injecting semantic ingredient information into the visual analysis pipeline. On the FoodSeg103 benchmark, the proposed method achieves state-of-the-art performance. Specifically, integrating LIM-Q into the Mask2Former decoder with a Swin-L image encoder yields a mean Intersection over Union (mIoU) of 55.0. LIM-F also demonstrates strong generalization and competitive performance, reaching an mIoU of 54.4 under the same model (Swin-L+Mask2Former). Furthermore, its applicability extends beyond Transformer-based decoders, as evidenced by an improvement from 47.7 to 49.8 mIoU when integrated into a CNN-based architecture. Notably, the improved segmentation accuracy is achieved with only a moderate (at most 3.8 GB) increase in the GPU memory consumption during training. Thus, the proposed approach offers a practical and scalable solution for fine-grained food understanding.

Large language model agents often spend substantial wall-clock time waiting for tool call results. Tool-call speculation can hide this latency by predicting and pre-executing an agent's next tool call if the prediction matches the agent's eventual tool call, but existing speculators are typically separate draft models or cached traces that are poorly aligned with the deployed agent's own behavior. We identify this speculator-agent gap and show that the target agent itself is a strong next-call speculator. This points to a simpler design: unifying the agent and speculator within the same model. In this paper, we introduce the self-speculating agent, a single model that both solves tasks in agent mode and predicts its next tool call from partial trajectories in speculator mode, fully reusing prefix KV cache. To enable this dual-mode agent without degrading performance, we propose a joint agent-speculator reinforcement learning method, which derives speculation targets from the agent's own rollouts and alternates agent and speculator updates. Across agentic search QA and conversational tool-use agentic tasks, our method improves average next tool-call Hit@1 from 44.1 to 61.2 for Qwen3-4B and from 48.9 to 66.3 for Qwen3.5-4B, while preserving agent task success.

Most video editing systems still lack explicit layered video representations, limiting their ability to perform realistic compositing, object reuse, and consistent manipulation. This limitation is especially pronounced in video object insertion and video layer decomposition, where existing methods rely on implicit inference or per-scene optimization due to the absence of explicit foreground-layer supervision. We introduce TriLayer, a large-scale triplet video dataset containing aligned composite, background, and foreground videos, where the foreground layers include both object appearance and associated visual effects. This explicit supervision enables models to learn layered video representations directly rather than inferring them implicitly. Building on this dataset, we propose DBL-Diffusion, a dual-branch diffusion framework that jointly models RGB composites and RGBA foreground layers through shared denoising and cross-branch interaction. We instantiate the framework in two tasks: DBL-Insert for layered object insertion, which generates explicit RGBA layers for realistic compositing and flexible post-editing, and DBL-Decompose for video layer decomposition, which recovers foreground and background layers using triplet supervision. Experiments demonstrate that explicit layer modeling substantially improves both insertion fidelity and decomposition quality.

An often overlooked factor of robot manipulation performance is the embodiment of the robot itself. Motivated by this problem, we study motion-conditioned robot co-design, where the goal is to generate complete robot designs that track target end-effector trajectories (from human demonstrations) while optimizing user-defined rewards. We introduce Transformer Transformer, a diffusion transformer trained on RoboTokens, a unified tokenization of robot embodiments, states, and actions. The same architecture can be used across embodiment spaces (e.g., wheeled bimanual, quadrupeds, humanoids) and use cases (embodiment generation, cross embodiment controller). Rather than overfitting to one reward function, Transformer Transformer is a dynamics model, whose reward-agnostic state and action predictions can be converted into reward-specific value predictions. These value predictions are used to steer embodiment diffusion towards high value robot designs, through a procedure we call Dynamics Self-Guidance. Experiments across multiple design spaces show zero-shot optimization of unseen rewards and trajectories, improving performance and runtime over the evolutionary baseline. Finally, we fabricated an optimized ALOHA design, which reduced tracking error by over 70% compared to the original design.

Sentimental Image Captioning (SIC) requires balancing emotional expression with visual fidelity. Existing methods often struggle with this trade-off, leading to hallucinations due to insufficient local grounding and the lack of sentimental verification mechanisms. To address these limitations, we propose SEA-Cap, a Sentiment-Evidence-Aware Multi-Agent System for faithful and evidence-grounded sentimental image captioning. SEA-Cap incorporates a Sentiment Evidence Miner that extracts structured, local affective cues to shift sentiment control from global attributes to verifiable object-level evidence. Leveraging this evidence, our framework orchestrates a collaborative workflow where a Generator, Hallucination Checker, and Arbitrator iteratively refine captions via a shared blackboard. By explicitly auditing generated content against mined visual evidence, SEA-Cap ensures both sentiment accuracy and factual consistency. Extensive experiments on two benchmark datasets demonstrate that SEA-Cap effectively mitigates hallucinations and achieves state-of-the-art performance.

Many ceiling construction tasks still rely on heavy serial manipulators that are difficult to deploy in cluttered interiors, motivating lightweight, field-ready alternatives that reach ceiling height while maintaining millimeter-level accuracy and the stiffness demanded by overhead tool loads. We introduce Tripody, a wheeled 3-DoF parallel robot for high-reach tasks that replaces the base spherical joints of a classical 3-SPR (3 legs; S: base spherical joint; P: actuated prismatic joint; R: end-effector revolute joint) morphology with universal joints, intentionally overconstraining the mechanism; small, distributed elastic deflections absorb the resulting incompatibilities, preserving predominantly translational motion. The 33kg system extends from 1.7m to 3.4m in height, supports a continuous 32kg payload, and offers a modular end-effector interface for ceiling operations. We detail the mechanical design - including custom linear actuators and a kinematic-compatibility analysis - and a control stack for accurate positioning that combines SE(3) state estimation, forward kinematics, and task-space control. In experiments, Tripody exhibits similar in-plane stiffness to a spherical-base variant but substantially higher torsional stiffness - an increase of 67% at 1.7m, 196% at 2.6m, and 454% at 3.4m - while maintaining negligible cross-axis coupling. Closed-loop positioning with a total station converges below 0.6mm across the entire workspace; pure model extrapolation achieves a 95th-percentile error of 2.7mm (max 3.6mm). Finally, we demonstrate task-level ceiling-drilling feasibility in an open-loop study by drilling a 15-hole pattern with 4.5mm maximum relative hole-position error after rigid alignment. These results support overconstrained, compliance-absorbing 3-SPR-like architectures as a practical path to lightweight, high- reach, millimeter-accurate construction robots.

Most existing drone-view geo-localization (DVGL) benchmarks contain drone imagery captured under a single illumination condition and lack geographically aligned visible drone images, infrared drone images, and satellite images from the same locations. To evaluate the generalization capability of DVGL methods under challenging illumination conditions, some methods train models on a visible benchmark and test them on an independent infrared benchmark. This protocol essentially constitutes transfer between datasets, which makes it difficult to systematically evaluate DVGL across daytime and nighttime conditions within a unified benchmark. To address this limitation, we construct IRCHN,a real-world DVGL benchmark designed for localization across different illumination conditions. IRCHN contains 26,460 images collected from 8,820 geographic locations across four representative scene categories, including farmland, coastline, forest, and urban areas. Each location provides one visible drone image, one infrared drone image, and one corresponding satellite image, which enables unified evaluation of DVGL methods across different illumination conditions and sensing modalities. We further propose the Modality-Adaptive State-Space Transport Relation Network (MASTR-Net), a DVGL framework tailored to localization under varying illumination conditions. MASTR-Net integrates modality-adaptive feature enhancement, bidirectional selective state-space relation modeling, and soft optimal transport relation alignment to jointly reduce modality gaps and view-induced structural discrepancies. Extensive experiments demonstrate that MASTR-Net outperforms existing state-of-the-art methods on IRCHN for localization under varying illumination conditions and achieves competitive performance on two infrared benchmarks, IR-VL328 and CVGL-RGBT. Code: https://github.com/SongtianhaoXu/MASTR-Net

Enterprise agents often need to integrate heterogeneous knowledge sources: documents for narrative facts, tables for computation, and dependency graphs for file relationships. Existing benchmarks typically evaluate retrieval or tool use without distinguishing whether an agent first selects the appropriate knowledge sources. We introduce WorkSurface-Bench, a benchmark for evaluating this capability as surface routing. It contains 1,151 atomic tasks derived from persona-scoped Workspace-Bench-Lite workspaces, spanning document, table, graph, and cross-surface questions. Its reference answers are auditable: table answers are reproduced through executed DuckDB queries, document answers are grounded in verified text spans, and graph answers are traced to source dependency annotations. We evaluate four model backbones across six controlled agent settings, yielding 27,624 protocol-error-free trajectories. Under gold-constrained tool access, agents achieve 98.7-99.8 Route F1, while Answer remains only 56.1-75.3 percent, showing that correct surface selection is necessary but insufficient for task completion. Matched interventions further show that surface hints improve Answer for three of four models, whereas removing irrelevant tools primarily improves routing and efficiency. In an independent three-annotator audit, all 200 sampled tasks pass all six quality criteria by majority vote, with 192 receiving unanimous judgments on every criterion. We release the dataset, construction pipeline, scoring code, and agent harness at https://github.com/haolpku/WorkSurface-Bench.

Cooperative navigation of multi-agent UAVs in complex environments faces key challenges including local optima traps, sparse rewards, learning imbalance among agents, and insufficient cross-scenario generalisation. This paper proposes a multi-agent deep reinforcement learning framework that addresses these issues through coordinated exploration, demonstration exploitation, safe curriculum scheduling, and structure-aware generalisation. First, a perception mechanism combining memory of visited states, directional novelty estimates, and penalty backpropagation enables agents to proactively detect and escape local optima. Second, a hierarchical collaborative demonstration buffer with tiered behaviour cloning manages trajectories by degree of team collaboration and applies differential supervision to the actor network, improving demonstration utilisation under sparse collaborative signals. Third, a safety-aware dual-condition curriculum scheduling mechanism reviews mastered scenarios through back-testing and experience pre-filling during training, suppressing catastrophic forgetting while ensuring both task performance and flight safety. For generalisation, local geometric features computed from sensor readings are abstracted into a domain parameter, through which a structure-aware gating network and mixture-of-experts mechanism condition the policy on local structural patterns rather than scenario-specific coordinates, enabling cross-scenario transfer without exposure to the target environment. The framework is further validated under mixed static-dynamic obstacle settings, showing robust adaptability to dynamic disturbances. Simulation results confirm strong performance in collaboration success rate, navigation robustness, zero-shot cross-scenario generalisation, and dynamic environment adaptability.

In this paper, we propose the Rule-based Visiting Circulation (RVC) model in tackling the challenge in the IEEE Big Data Cup 2022: Trip Destination Prediction. Given trips containing travel information, personal attributes, origin zones, and their features in the training metropolitan areas, the task is to predict the destination of every trip in a targeted metropolitan area whose destinations are not given at all at the training stage. We highlight the challenges in this destination prediction task -- having no knowledge of the destinations in the targeted metropolitan area. We provide insights from the datasets, in which revisiting behaviors and the relationships between origins and destinations play a crucial role in individuals' trips. Hence, we design a simple but comprehensive method, rule-based visiting circulation, which directly utilizes the origin information and individuals' trip behaviors to determine the destinations in the targeted metropolitan area, i.e., requiring no learning from the four training areas. Experimental results on both offline evaluation and leaderboard submission consistently exhibit the proposed RVC can significantly outperform supervised learning methods and other heuristics. The RVC method eventually brings us to second place in the competition leaderboard.

Bimanual teleoperation provides an effective way to collect robot demonstrations, but it assumes that the operator and robot have matching numbers of simultaneous control channels. This assumption breaks for tri-manual systems: the robot can coordinate three arms concurrently, whereas a single operator can continuously control only two. Pairwise mode switching may therefore record otherwise independent motions sequentially, causing behaviour cloning to reproduce delays imposed by the interface rather than required by the task. We present TriManPolicy, a tri-manual imitation learning system that allows one operator to demonstrate behaviours for three arms. Its central component is Dependency-Aware Tri-Arm Scheduling (DATS). The key idea is to preserve the demonstrated arm motions while reconsidering when they occur. DATS retimes demonstrations offline by preserving local sensorimotor segments of fixed duration and repositioning them according to constraints on task order and arm usage that are reviewed by a human. The resulting data train a single synchronous policy for all three arms, while deployment requires neither the dependency graph nor the scheduler. Across six challenging tasks performed in the real world, policies trained on demonstrations retimed by DATS exhibit more efficient coordination while maintaining comparable observed task success. Offline analysis further shows that DATS changes the supervision across arms rather than merely removing idle periods.

This paper presents a cooperative indoor UAV guidance framework that combines a shared voxel-map world model with a multi-agent Soft Actor-Critic (MASAC) controller. Multiple drones fuse 360 LiDAR observations into a common world-frame occupancy map, which is converted into a compact bird's-eye-view (BEV) representation and provided to each agent as an ego-aligned local crop. This integrate-in-world, act-in- ego design enables consistent multi-UAV spatial fusion whilst retaining decentralised continuous control. The policy combines BEV map features, near-field obstacle observations, and compact goal and peer-state information within a centralised-training, decentralised-execution framework. In simulation, the learned controller achieves a 90.3% success rate in corridor navigation, outperforming Astar planning, an artificial potential field controller, and a prior guidance method. To address residual sim-to-real mismatch, the simulation-trained policy is further adapted using offline imitation fine-tuning from real-world data. Real-world experiments in GNSS-denied indoor environments demonstrate stable two-UAV cooperative operation across increasingly chal- lenging obstacle layouts. The results show that shared voxel-map representations provide an effective and scalable spatial substrate for learned cooperative indoor UAV guidance.

Large language model (LLM) agents increasingly rely on invoking external tools to complete real-world tasks. Tool retrieval, which selects a small task-relevant subset from a library of thousands of tools before the agent acts, has therefore become a critical component of LLM agent pipelines. However, existing retrievers either score each tool in isolation or assemble the tool set sequentially, so the joint utility of a candidate set is never evaluated as a whole. In this paper, we propose HYSET, short for HYperedge-based SEt-level Tool retrieval. Our contributions are threefold: (i) we formulate tool retrieval as query-conditioned hyperedge prediction on a tool co-invocation hypergraph, under which the tool set itself becomes the unit of scoring and most existing retrieval paradigms reduce to restricted instances; (ii) we capture size-dependent tool compatibility through cardinality-specific interactions; and (iii) we design HYSET as a pre-selection module requiring no modification to the downstream agent. Experiments on ToolBench demonstrate that HYSET consistently outperforms state-of-the-art baselines in both tool retrieval performance and end-to-end task success. Beyond the in-domain setting, HYSET further supports zero-shot/few-shot transfer, generalizing to held-out tools/categories and unseen domains with minimal supervision.

Fall represents a significant risk of accidental death among individuals aged over 65, presenting a global health concern. A fall is defined as any event where a person loses balance and moves to an off-position, which may or may not result in an impact where the person hits the ground. While fall detection systems have achieved good results in general, impact detection within falls remains challenging. This study proposes an efficient methodology for accurately detecting impacts within fall events by incorporating 3D joints skeleton data treated as a graph using Spatio-Temporal Graph Convolutional Networks (STGCN), Gated Recurrent Unit (GRU), and Bidirectional Long Short-Term Memory (BiLSTM) layers. By pinpointing impact moments, our approach enhances precision by distinguishing between false falls and actual impacts, contributing to better healthcare resource allocation. Our methodology, evaluated using the improved 3D skeletons UP-Fall dataset, achieves accuracy exceeding 90\% across various fall scenarios. We have made this improved dataset publicly available at https://zenodo.org/records/12773013 to facilitate further research.

Full-field reconstruction of air pollution is essential for evaluating pollution exposure and supporting public health decision-making. However, the complex interactions among pollutants, hard-to-predict weather patterns, and limited monitoring station coverage make this a complex task. We apply deep learning techniques to provide fast and accurate reconstructions from sparse observations of four key pollutants: NO2, O3, PM2.5 and PM10. Models are trained on full-field simulation data and evaluated on real-world observations collected from 9 to 28 monitoring stations in the city of Paris. We introduce a diffusion-based generative framework for multi-pollutant reconstruction and benchmark its performance against deterministic deep learning models. Despite noisy observations and strong spatial variability, the models achieve high structural similarity on simulated validation data and produce realistic spatial patterns on real-world observations, as indicated by power-spectrum analysis. We introduce data augmentation methods that enable transfer to real-world observations without retraining, allowing the models to generalise beyond the training period. These findings highlight the potential of ML models for reliable real-world deployment in air pollution reconstruction tasks.

Cognitive distortion amplifies negative emotions and contributes to mental health disorders. Cognitive Behavioral Therapy (CBT) is an effective way to address cognitive distortions, but its large-scale application is limited by the shortage of professional therapists. Although large language models (LLMs) have recently been explored for mental health applications, existing methods still suffer from limited domain specificity, overly flattering responses, and the absence of well-defined annotations for cognitive distortions. This paper proposes Cognivia, an evidence-based artificial intelligence therapist that integrates automatic cognitive distortion identification and rational response generation. Our framework is built on authoritative CBT texts widely regarded as core paradigms and standard references. It is further augmented with mental health question-answer (Q and A) data, and employs multi-stage prompting and structured generation strategies under the supervision of behavioral science experts. Then we fine-tune a lightweight LLM on this augmented CBT dataset to obtain Cognivia. In addition, we propose the first hierarchical quality evaluation framework for assessing LLM-generated rational responses, developed through collaboration between AI researchers and behavioral science experts. Cognivia is evaluated using lexical metrics, LLM-based Judges with two complementary criteria, and human evaluation by 10 behavioral science experts. It consistently outperforms the baseline methods in cognitive distortion recognition and rational response generation, demonstrating its effectiveness. Our code is available at https://github.com/SNOWTEAM2023/Cognivia.

Text-space optimization adapts large language models (LLMs) by editing external natural-language artifacts rather than model weights, so the optimized artifacts remain inspectable and the model can be treated as a black box. However, most existing text-space methods keep evaluation fixed. On open-ended tasks, this can become a bottleneck: once the solver improves on the criteria a rubric measures, omitted dimensions remain invisible to the optimization signal. Simply evolving the rubric is also unreliable when updates are selected by the current solver's score, because apparent progress can come from making the rubric easier to satisfy. We introduce DecoEvo (Decoupled Co-Evolution), which co-evolves a solver skill and a rubric-generator skill under decoupled objectives without using gold rubrics during optimization. The solver skill is updated using criterion-level feedback, while the rubric-generator skill is revised through complementary audits of requirement coverage and response discrimination that are independent of aggregate solver score. This separation focuses generator updates on newly exposed solver weaknesses, reducing repeated emphasis on criteria the solver already satisfies. Under each benchmark's official evaluation, DecoEvo outperforms all compared methods across five benchmarks and three LLM backbones, yielding 2.8--5.0\% relative gains over SkillOpt in the five-benchmark average.

Emerging Omni-modal Large Language Models (OmniLLMs) enable unified understanding of text, audio, and video, but their long audio-video token sequences introduce substantial memory and inference costs. Existing compression methods mainly focus on selecting important tokens under fixed budgets, leaving the preceding budget-allocation problem underexplored. We show that direct query-to-audio/video similarity is unreliable for inter-modal budget allocation, and that uniform intra-modal budgets can miss key evidence while retaining redundant content. To address these limitations, we propose OmniDelta, a training-free, skill-driven framework that couples intent-aware inter-modal allocation with content-aware intra-modal allocation. OmniDelta first constructs audio and video skill pools to shift the fixed retained-token budget according to query demand, then reallocates modality budgets over audio segments and video frames using local complexity and temporal redundancy. The resulting local budgets can be combined with existing pruning strategies, preserving the total retained-token ratio while changing where the budget is spent. Experiments on four audio-video benchmarks with two Qwen2.5-Omni models show that OmniDelta establishes a new accuracy-efficiency Pareto frontier across pruning ratios. At 25% token retention on Qwen2.5-Omni-7B, OmniDelta reduces GPU memory by 22.0% and achieves a 1.64x end-to-end speedup over full-token inference.

Transformer adaptation is typically distributed across model depth, even when the intended change is narrow. We investigate how adaptation site shapes what a model learns, how well that learning generalizes, and how selectively it is applied. We introduce a controlled benchmark spanning five objectives (lexical binding, factual association, behavioral policy learning, causal mapping, and procedural reasoning) and define each objective's "adaptation geometry" as its profile of acquisition, transfer, and boundedness under full-stack and early-, middle-, or late-layer LoRA. The objectives exhibit distinct geometries. Lexical binding favors early-layer adaptation for acquisition and boundedness but requires broader updates for transfer; factual association favors later layers among localized adapters; behavioral learning separates late-layer action acquisition from middle-layer policy gating; and causal and procedural transfer benefit most from middle- or full-stack adaptation. These patterns largely persist under parameter-matched controls, and most corresponding directional contrasts replicate across five model families. These findings establish adaptation site as a key design variable for controlling what models learn, generalize, and leave unchanged.

Complex tasks often decompose into parallelizable yet interdependent subtasks, making orchestration critical to the performance of multi-agent systems (MAS). Existing evaluations typically rely on end-to-end execution, which conflates orchestration-plan quality with worker capabilities, tool reliability, and environmental noise. Moreover, the time and token costs of real execution grow rapidly with workflow scale, making systematic evaluation expensive. We present OrchBench, a simulation-based benchmark for evaluating multi-agent orchestration plans in isolation. Starting from real-world tasks, OrchBench constructs directed acyclic graphs (DAGs) that encode task dependencies, with controlled sizes and degrees of parallelism. Given a DAG, a per-agent context limit, and an agent budget, the evaluated planner assigns subtasks to agents and specifies cross-agent information transfers and their retention ratios. A deterministic simulator evaluates the resulting plan without invoking worker agents and returns interpretable measures of result quality, makespan, and token cost. The simulated scores produced by OrchBench correlate strongly with quality scores from Claude Code executions, achieving a Pearson correlation of \(r=0.816\), while requiring only \(1.3\%\) of the tokens and \(10.3\%\) of the wall-clock time. Across diverse planners and workflow scales, we find that preserving task-critical information is more important than simply increasing the number of agents, and the benefits of parallelism diminish as coordination failures accumulate. These results establish OrchBench as an efficient and interpretable benchmark for comparing and diagnosing multi-agent orchestration plans.

Public services face growing pressure to adopt artificial intelligence (AI) to close the gap between rising demand and falling resources. That pressure has intensified with general-purpose AI (GPAI): AI built on large language models that can be directed by prompt alone to perform an effectively unbounded range of tasks. We argue that the properties that make these models attractive - their generality, accessibility, and low deployment cost - undermine the conditions under which AI safety has historically been pursued. The safety concepts that public service governance frameworks foreground - accuracy, bias, explainability, and accountability - were made tractable by narrow, purpose-built AI, and the mitigations that guidance documents prescribe presuppose exactly what GPAI removes. Accuracy cannot be quantified over unbounded outputs. Bias cannot be disaggregated when outputs are free-text judgements rather than categorical predictions. Explainability gives way to the appearance of explanation, and accountability erodes as outputs are optimized to persuade. We develop this through the case of policing, where the consequences of governance failure are most severe, and show why the same failure is likely to recur across other public services. The two mitigations that dominate policing AI strategy - expert evaluation and human-in-the-loop oversight - both rest on assumptions that GPAI violates. Safety assurance thus shifts from an intrinsic feature of building an AI tool to an optional add-on. We recommend a clear taxonomic distinction between narrow and general-purpose AI in governance documentation, a preference for technological parsimony, a pause on operational deployment of GPAI in policing until adequate evidence exists, and a coordinated national safety infrastructure with the authority to generate that evidence and determine when responsible deployment is achievable.

Instruction-based Image Editing (IIE) aims to transform a given image into a new one based on textual instructions. Advances in Large Language Models (LLMs) and Vision-Language Models (VLMs) have accelerated progress toward practical ``one-sentence image editing" systems. This survey presents a systematic taxonomy and comprehensive review of IIE research, structured around five core dimensions: (1) task definition and hierarchical categorization of editing operations, (2) methodologies for training data construction, (3) architectural evolution from GAN-based to diffusion and autoregressive paradigms, (4) standardized evaluation metrics and benchmark development, and (5) introduction of commercial solutions. Our analysis shows critical technological milestones across model generations. We further propose a Comprehensive, in-Depth, and Diagnostic benchmark for IIE task (CDD-IIE Bench), which can rigorously assess the multiple aspects of model performance. Through empirical comparisons of open-source solutions, we highlight their respective capabilities and limitations. Finally, we discuss future research directions to advance the field.

While text-to-image models exhibit remarkable visual fidelity, they frequently violate fundamental physical commonsense. Existing benchmarks often rely on coarse-grained descriptions, failing to diagnose the mastery of specific physical principles. Moreover, the high stochasticity of generative processes causes current prompt optimization methods to suffer from gradient hallucinations, where optimizers are misled by transient visual artifacts rather than systemic flaws. To address these challenges, we introduce OmniPhys, a rigorous benchmark of 1,551 samples grounded in a Physical Knowledge Graph. By aligning PhET simulations with standard curricula, OmniPhys operationalizes a knowledge-to-scenario pipeline that performs diagnostic stress tests via a dual-path verification protocol. We further propose OmniPrompt, an iterative framework that treats physical alignment as a discrete optimization problem. For each query, OmniPrompt aggregates K stochastic images into a per-query feedback buffer. Across training, it further merges feedback from batches of B queries before each meta-policy update, filtering seed and query-local noise. Evaluations across 12 representative text-to-image models reveal universal physical bottlenecks. Results demonstrate that OmniPrompt significantly enhances physical consistency across diverse backbones, proving the transferability and efficacy of our evolved meta-policies. The code and data are available at https://github.com/zjukg/OmniPhys

F(AI)2R is FAIR research with AI in the loop, twice: an AI-assisted authoring pass and a machine-readable audit pass over every artefact. AI systems now draft, refactor, and verify research artefacts, yet their contributions are rarely recorded in a form a later human or machine can audit. Building on the original F(AI)2R experiment, we generalize its provenance model beyond scholarly writing into aiprov, a PROV-O extension covering any AI-in-the-loop artefact, and we package the method as an executable skill that an AI agent operates itself: setup asks the human operator for their ORCID ID, resolves their identity from the public registry, and scaffolds continuous integration that gates every push on graph conformance and publishes the current build of this very paper. The paper is its own case study. Every activity, claim, and source in its production is recorded in the repository's provenance graph under two invariants: no parentless claim, and verification rungs that only humans may grant.

We present AIriskEval-edu Demo, a platform that audits the pedagogical quality of instructional explanations and provides explainable audit results. The platform evaluates an explanation against a rubric covering five dimensions of pedagogical risk: factual accuracy, depth and completeness, focus and relevance, student-level appropriateness, and ideological bias. For each dimension, it returns a binary decision and a confidence score. Detected risks also include a natural-language rationale and, except for Depth and Completeness, a localized evidence span. The platform integrates GPT-5.5 through an external API and a self-hosted Llama 3.1 8B evaluator that runs on consumer-grade GPUs. The local evaluator is fine-tuned on AIriskEval-edu, a dataset of K-12 instructional explanations with risk and explainability annotations. The platform operates in two modes: in AI mode, both evaluators assess stored explanations generated under six simulated teacher profiles, each representing a distinct pedagogical behavior and potential risk; in human mode, the local evaluator audits user-written explanations in real time. The local evaluator outperforms GPT-5.5 on most reported metrics, offering educational institutions a practical way to keep audited content within their own infrastructure.

Interdisciplinary research is accelerating, yet scientific papers remain difficult to understand outside their home fields. We study large language model (LLM)-based simplification of scientific texts and present a human-in-the-loop workflow that transforms expert summaries into more accessible versions for non-specialists. Using SciSummNet as the source corpus, we first generate baseline simplifications with GPT-4o-mini. In Phase 1, readers from STEM fields outside computer science identify difficult sentences and phrases and compare the original and GPT-simplified summaries in terms of comprehensibility, naturalness, and simplicity. In Phase 2, computer science experts use this feedback to create expert-edited reference simplifications. We release the resulting corpus together with human judgments and automatic evaluation results. The Phase 1 judgments show a clear preference for the GPT-generated summaries in terms of comprehensibility and simplicity, while qualitative analysis of the Phase 2 edits highlights the importance of preserving domain-specific terminology and the strength of scientific claims. The resulting resource supports the training and benchmarking of simplification systems for cross-disciplinary scientific communication.

Personalized person image generation requires preserving subject identity across both local facial details and broader appearance cues. Existing methods typically emphasize only one level of identity information, leading to an inherent trade-off between facial fidelity and overall appearance consistency. To address this, we first propose a simple dual-branch baseline that unifies global appearance control and local facial control within a shared generation framework. This simple combination of different branches yields promising results, but suffers from instability in practice due to uncoordinated branch contributions. To this end, we propose Dynamic Balancing Scaling (DBS), a fine-tuning strategy for improving face and appearance identity coordination. DBS consists of two components: adaptive temporal gating, which dynamically modulates branch contributions along the denoising trajectory, and region-aware optimization, which improves the coordination of facial, appearance, and global supervision. Together, these designs alleviate persistent face-branch over-dominance and encourage more effective appearance-aware guidance. We also introduce Pexels-100, a benchmark for evaluating holistic identity consistency in personalized person generation. Experiments show that DBS achieves a better trade-off between facial fidelity and appearance consistency than existing open-source baselines, while providing a controllable basic framework for holistic identity modeling.

Quattrociocchi and colleagues warn that the fluent outputs of large language models may allow linguistic plausibility to substitute for epistemic evaluation, producing the condition they call *Epistemia*: the experience of possessing knowledge without undertaking the practices through which judgment would ordinarily be warranted. This article accepts that diagnosis but challenges its explanatory framework, which compares an embodied, socially situated human knower with an isolated generative model thereby locating epistemic legitimacy in capacities internal to autonomous agents. Drawing on Carlo Sini's philosophy of practices, writing, signs, and technics, we propose instead to understand a large language model (LLM) as a *techno-semiotic machine* that automates a phase of written semiosis by producing plausible linguistic configurations from the sedimented archive of human writing. From this perspective, *Epistemia* is one consequence of a broader phenomenon that we call *epistemic schizologia*: the socio-technical cleavage between signs as linguistically accomplished expressions and signs as moments within socially embedded circuits of interpretation, evidence, criticism, verification, and responsibility. This cleavage is reinforced by *eikotic closure*, through which a plausible continuation is presented with the finality of an epistemic result, and by algorithmic authority and epistemic self-misrecognition. The relevant unit is therefore not the model alone but the complete practice in which generated inscriptions are prompted, interpreted, verified, contested, used, and made consequential. This reframing preserves the distinction between linguistic production and responsible understanding while grounding a design programme centred on inspectable genealogy, contestability, distributed responsibility, epistemic agency, and the evaluation of hybrid human--AIpractices.

Adapting Large Language Models (LLMs) to specialized domains often incurs an alignment tax, as fine-tuning on domain-specific tasks can cause catastrophic forgetting and substantially degrade performance on general tasks. We propose MemSFT, which mitigates the alignment tax by decoupling domain specialization from backbone parameter updates through a plug-and-play parametric memory. The memory is trained to imitate the behavior of a non-parametric retriever operating over domain data, thereby memorizing knowledge and patterns that would otherwise be accessed through retrieval. Once trained on a specific domain, the memory can be reused across LLMs of different sizes. During generation, a learned router dynamically fuses the output distributions of the memory and backbone at each decoding step, allowing domain expertise to be invoked selectively. Across biology, geoscience, and law, evaluations with models ranging from Qwen3-8B to Qwen3-235B-A22B show that MemSFT consistently improves domain performance with negligible degradation in general performance, whereas full SFT suffers severe forgetting on general tasks. Overall, our results demonstrate a practical path to decoupling general model capabilities from domain-specific knowledge at the parameter level, thereby equipping LLMs with new specialized capabilities without compromising their general capabilities.

Physics-informed neural networks (PINNs) have emerged as a powerful paradigm for solving partial differential equations (PDEs) by embedding governing physical laws into deep neural networks. However, their reliance on computationally expensive gradient-based optimization and deep architectures often results in slow training, high computational cost, and limited scalability. In this work, we propose a novel physics-informed broad learning system (PI-BLS), the first physics-informed learning framework based on broad RdNNs. The proposed formulation embeds the governing differential operator and the associated initial and boundary constraints directly into a linear output-layer optimization problem, thereby replacing nonlinear gradient-based training with a deterministic least-squares solution obtained via the pseudoinverse. Consequently, the entire learning process is reduced to a single linear optimization stage while preserving the underlying physical constraints. As a result, PI-BLS offers an efficient learning paradigm for a physics-informed learning framework for solving PDEs that eliminates iterative backpropagation while preserving the underlying physical constraints. Experimental results on representative forward PDE benchmarks demonstrate that PI-BLS achieves competitive and often superior performance with reduced training time and model parameters compared with conventional PINNs.

Retrieval-augmented generation improves knowledge-intensive question answering, but indiscriminate retrieval can introduce irrelevant evidence and unnecessary computation. We investigate whether verbalized confidence from black-box language models can serve as an actionable signal for retrieval routing. Our method, BeyondUncertainty, first elicits a structured provisional answer and confidence estimate, then applies a model-specific threshold selected on held-out validation data and frozen before test evaluation. Low-confidence questions receive top-$5$ TF--IDF retrieval followed by a second answer call, whereas high-confidence questions return the provisional answer directly. We evaluate 27,000 policy instances across six QA benchmarks, three model families, and three retrieval policies. BeyondUncertainty achieves $0.483$ mean token-level F1, compared with $0.467$ for always retrieval and $0.401$ for no retrieval, while reducing retrieved passages by $20.4\%$ relative to always retrieval. When matched on the number of questions routed to retrieval within each dataset--model cell, it outperforms a post-hoc random allocation in 17 of 18 settings, with an average gain of 0.024 F1. Although poorly calibrated as an absolute probability, probe uncertainty modestly predicts question-level retrieval benefit (AUROC $=0.628$). However, the additional probe increases total token usage by $28.2\%$, revealing a trade-off between more selective evidence acquisition and end-to-end token efficiency.

Robotic hardware evolves over time, but demonstration data is often tied to a specific sensor and actuator configuration. This raises a practical and underexplored question: when does legacy data begin to benefit an upgraded robot? We study this question on a wheeled humanoid platform across two hardware generations, where both the camera and gripper are changed while the overall morphology remains fixed. Contrary to the common assumption that more cross-configuration data is always helpful, we observe a grokking-like transition: legacy data remains ineffective until the upgraded configuration acquires a minimum level of task competence, after which co-training gains rise sharply before diminishing near saturation. We hypothesize that this task-dependent transition is governed by a transfer threshold and characterize the resulting three-phase pattern. Across real-robot manipulation tasks, we observe all three phases: no measurable benefit at low competence ($10.0\% \rightarrow 10.0\%$), a sharp gain after crossing the threshold ($23.3\% \rightarrow 86.7\%$ on flower insertion), and diminishing returns at high competence ($85.0\% \rightarrow 93.3\%$ on pen insertion). We provide a theoretical account based on gradient alignment and residual policy uncertainty, and derive a phase-aware rule for deciding when to collect more new-hardware data and when to reuse legacy demonstrations. We further validate this three-phase pattern on a mobile dual-arm watering task, with results consistent with our predictions.

Medical-imaging AI benchmarks combine datasets, DICOM rendering, prompts, provider APIs, automated labels, statistical code, manuscripts, and repository releases. Agreement across these artifacts is usually assumed rather than tested. We performed a retrospective forensic reproducibility audit of a preserved chest-radiograph vision-language model (VLM) pilot; no model was called again and no image or report was newly annotated. We traced prompt bindings, DICOM metadata, output completeness, label extraction, matched analyses, and release propagation. Of 300 planned model-prompt calls, 297 yielded nonempty reports. Sixty Claude calls labeled A/B were executed with the same C prompt. The 30 studies represented 28 patients. Four MONOCHROME1 images were rendered without required polarity inversion, dataset split membership was not retained, and the unvalidated extractor truncated five reports to 4000 characters. Reconstructing one common cohort of 369 complete case-finding blocks changed Cochran's Q from 154.73 to 182.29. Of 45 McNemar comparisons, 27 had unadjusted p < 0.05 and 20 remained below 0.05 after Holm adjustment. These values describe only the archived automated-label matrix; they do not recover the intended prompt comparison or establish clinical performance. We withdraw the original performance, ranking, prompt-effect, and clinical claims and specify machine-verifiable controls for cohort, DICOM rendering, prompt and model identity, call status, annotation provenance, keyed analysis, and derived artifacts.

This paper proposes a system-aware adaptive channel state information (CSI) feedback framework for massive multiple-input multiple-output (mMIMO) systems, aiming to dynamically optimize the trade-off between reconstruction fidelity and signaling overhead. While deep learning-based autoencoders (AEs) have enabled significant CSI compression, conventional fixed-ratio schemes fail to adapt effectively to non-stationary channel conditions. To address this limitation, we develop a reinforcement learning (RL)-driven control framework that operates over a bank of pretrained multi-rate AEs, each corresponding to a distinct compression ratio (CR). At each time step, a centralized RL agent selects the most suitable CR for each user based on observed channel conditions and system performance indicators. Distinct from conventional mean squared error (MSE)-centric designs, we introduce a system-aware reward formulation that jointly accounts for spectral efficiency via signal-to-interference-plus-noise ratio (SINR), feedback overhead constraints, and the computational cost of model adaptation. Simulation results on high-dimensional delay-domain CSI datasets demonstrate that the proposed RL-guided framework effectively balances the overhead-accuracy tradeoff and adapts to dynamic channel environments. The proposed method improves spectral efficiency and feedback efficiency compared with fixed compression schemes and adaptive baselines, while maintaining a modest computational and memory footprint. Averaged over different numbers of users and across all considered baselines, the proposed RL framework reduces the CSI feedback cost by more than 53.4%, improves the average downlink sum rate by 53.64%, and reduces the NMSE by 22.38%. These results demonstrate its ability to achieve a more efficient rate-accuracy-feedback tradeoff under dynamic wireless conditions.

Parameter-efficient fine-tuning (PEFT) and low-bit quantization are now standard tools for adapting language models under tight compute budgets, yet their interaction is most often studied on billion-parameter models where the design space is expensive to explore. We ask a complementary question: on a specific, fully reproducible 60M-parameter encoder-decoder model (T5-small) and a single-table text-to-SQL benchmark (WikiSQL), how much task accuracy does each efficiency knob actually cost? We run a controlled, single-variable study over (i) LoRA rank r in {2, 4, 8, 16, 32}, (ii) the set of adapted modules, and (iii) numerical precision. We report task accuracy alongside system-level metrics including trainable parameters, peak training memory, inference latency, and throughput, and frame adaptation as a constrained trade-off rather than an accuracy-only objective. Our results show that LoRA with r=16 recovers within 11.6 percentage points of full fine-tuning accuracy (59.6% vs. 71.2% exact-match) while training fewer than 1% of parameters and consuming 31% less peak GPU memory. Within this setting, rank beyond r=16 yields no measurable accuracy gain. QLoRA with INT8 and NF4 quantization achieves comparable accuracy (52.8% and 53.2%) at dramatically lower memory cost (0.60 GB each), demonstrating a compelling trade-off for memory-constrained deployments. All code, configurations, and logs are released for full reproducibility.

Entity alignment (EA) identifies entities across knowledge graphs (KGs) that refer to the same real-world object. Conventional EA methods mainly exploit explicit graph structures and textual fields, which often provide insufficient semantic understanding to recognize the same entity under heterogeneous descriptions and distinguish it from semantically similar entities. Although large language models (LLMs) offer deeper entity understanding, existing LLM-based EA methods largely use this capability for auxiliary generation or candidate-conditioned decisions. Consequently, such understanding is not distilled into a stable and directly comparable identity space, leaving alignment tied to specific KG pairs or candidate sets and requiring repeated processing as the matching context changes. To address these limitations, we propose IRIS (Identity Representations from Internal States), a training-free framework that constructs for each entity an iris-like signature encoding its distinctive and stable identity characteristics. IRIS derives these signatures by eliciting identity-oriented contextual representations from a frozen LLM, thereby forming a shared space in which each entity is encoded once and can be aligned across different KGs through direct similarity comparison, without pair-dependent representation construction or candidate-wise LLM inference. Across four established EA benchmarks and two frozen LLM backbones, the best IRIS variants achieve Hits@1 scores of 100.00, 99.38, 98.31, and 97.99 on D-Y-15K V2, DBP-WIKI, ICEWS-WIKI, and ICEWS-YAGO, respectively.

Photoacoustic tomography (PAT) combines the optical absorption contrast of biological tissue with the spatial resolution of ultrasound, yet recovering the initial pressure distribution from sparse-view sensor measurements remains an ill-posed inverse problem. Iterative compressive-sensing solvers and unrolled deep networks both retain a dependence on the system matrix at inference, which leaves real-time clinical reconstruction computationally expensive. This paper proposes the Sensor Attention Network (SAN), a Transformer-based architecture that treats the full time series of each sensor as a token and maps raw measurements directly to the reconstructed image without invoking the system matrix at inference. For training and benchmarking, an analytical k-space H-matrix is constructed and validated against the k-Wave pseudo-spectral solver under matched geometry, achieving a mean per-sensor Pearson correlation of 0.919 +/- 0.049, with k-space apodization and Gaussian temporal damping acting synergistically to reduce the energy-normalized mismatch by 49%. Trained with a vessel-weighted loss on 488 augmented samples and evaluated on 46 held-out samples against ISTA, split-Bregman total variation (SBTV), and learned ISTA (LISTA), SAN attains the highest mean SSIM (0.522) and PSNR (22.09 dB) and the lowest NMSE (0.233). Paired t-tests and Wilcoxon signed-rank tests confirm the superiority of SAN over LISTA on PSNR, NMSE, and Pearson correlation at p < 1e-8, and over ISTA and SBTV on all fidelity metrics. By bypassing the H-matrix at inference, SAN reduces reconstruction time by at least an order of magnitude, supporting real-time PAT reconstruction.

Recovering an editable design file from a raster image is a common and costly bottleneck in modern design workflows, yet remains challenging since editability depends on recovering multi-modal attributes, such as typography, vector geometry, colors, grouping, and layer ordering. We present ReDesign, an agentic framework that grows an editable layer hierarchy by selecting and composing specialized tools across modalities. To keep this long decision process reliable despite imperfect tool outputs, we introduce graceful verification at each expansion, which provides local accept, prune, or retry feedback that prevents error accumulation and avoids large scale reruns. To evaluate editability at scale, we introduce the Figma Edit Replay Benchmark, consisting of 909 raw Figma files and 14,796 controlled edit instructions that replay edits on reconstructed outputs. Across this benchmark and standard reconstruction metrics, ReDesign achieves strong visual fidelity while delivering the highest editability across layout, color, and text edits, outperforming layered decomposition baselines and serial tool use pipelines.

This paper proposes a differentially private distributed cooperative control scheme for multi-agent systems (MAS). Unlike conventional approaches that actively inject artificial noise for privacy protection, this work investigates whether inherent communication noise can itself serve as a natural privacy mechanism. A physically motivated communication-noise model is developed for mobile MAS by incorporating transmitter perturbation, receiver noise, path-loss attenuation, and log-normal shadowing. The resulting effective noise variance depends on inter-agent state differences, thereby capturing the distance-dependent signal perturbation arising in practice. Based on this model, a distributed finite-horizon Linear Quadratic Regulator (LQR) mechanism is designed to achieve formation tracking while protecting agents' private control preferences. Rather than protecting the full local cost function, the proposed privacy formulation focuses on the ratio of the LQR weighting matrices, which captures the trade-off between tracking accuracy and control effort when the quadratic cost structure is publicly known. A set-theoretic sensitivity analysis shows that this weighting-ratio adjacency formulation yields less conservative privacy bounds than gradient-based protection under the considered addition/removal adjacency relation. Theoretical analysis demonstrates that, under suitable design conditions, the proposed mechanism provides bounded cumulative (ε,δ)-differential privacy guarantees for the weighting ratios over an infinite horizon without artificial noise injection. Meanwhile, the cooperative tracking error is shown to converge almost surely and in mean square to a finite random limit, with its expectation remaining bounded. Numerical examples validate the theoretical results and illustrate the resulting privacy-performance trade-off.

Open-vocabulary segmentation labels arbitrary categories from a text query without per-class training, yet on remote sensing imagery it underperforms on categories it handles reliably elsewhere. We find that much of this gap traces to the text query rather than to the segmentation model. Because these models are not specialized for overhead imagery, the class name that serves as the query is often a weak address into the vision-language embedding space. We show that a better name repairs part of the gap, while the remaining failures call for an address that the tested natural-language rephrasings do not provide. We recover that address from a few examples through textual inversion on a frozen model, keeping inference text only. On a representative benchmark this raises the mean intersection over union on the affected categories from 3.9 to 39.4, and across eight remote sensing datasets it improves over few-shot methods that instead inject visual prompts at inference.

Agent skills package reusable procedures that improve downstream performance. Their lightweight, portable form enables marketplace monetization and private deployment behind cloud-hosted agent interfaces, giving providers incentives to keep high-value skills proprietary. Yet hiding the artifacts does not conceal their behavioral effects, which remain observable in execution trajectories and form a behavioral side channel. We define this exposure as Skill Leakage: reconstructing proprietary skills from trajectories elicited by benign queries, without reference answers or success labels. We introduce SigLeak, a black-box framework that exploits recurring skill signatures in agent behavior. It constructs diverse, decision-rich diagnostic tasks, contrasts matched skill-enabled and skill-disabled trajectories, and iteratively refines a reconstructed skill from the isolated patterns. Across five scenarios, three model families, and three agent frameworks, SigLeak outperforms or matches three baselines in nearly every setting. It raises the success rate by 6.88 percentage points over the skill-disabled reference on average and achieves the highest overall SkillSim, our metric for coarse- and fine-grained semantic similarity. These results show that benign execution trajectories can expose proprietary procedural knowledge. The code is available at https://anonymous.4open.science/r/SigLeak-D1DB.

Future event prediction carries broad social impact yet remains challenging. SOTA approaches augment LLMs with external agent frameworks whose predictive capability vanishes once the harness is removed. While recent Tool-Integrated Reasoning (TIR) internalizes deep search for multi-hop retrieval of facts, forecasting further demands temporal search and reasoning over historical trends and dynamic shifts. The key obstacle is data: historical queries induce temporal leakage that degrades forecasting into retrieval. Prior works either freeze information gathering with static observations, or rely on rejection sampling or unresolved fresh queries that discard vast amounts of data, degrading synthesis efficiency. We propose a time-truncation harness that enforces a temporal cut-off at every turn, enabling TIR-style sampling from historical events, reducing temporal leakage and reliance of rejection sampling or unsolved queries, increasing the sampling efficiency. We further build a large-scale corpus and a process-based metric and show that our harness naturally induces a broader temporal breadth of search and raises the proportion of high-quality data, further increasing the efficiency and reducing the reliance on complex rubrics. Distillation experiments show that students trained on harness-intervened data achieve the best performance, demonstrating harness-assisted model evolving that turns higher quality temporal search and reasoning data into a parametric advancement of the students.

Variational Autoencoders are widely used to encode high-dimensional and noisy observations in robotics. However, their stochastic latent creates a mismatch with Proximal Policy Optimization (PPO): an effective policy marginalizes over the latent distribution, whereas former implementations estimate its probability ratio and KL divergence using only one latent sample. We identify a fundamental but overlooked theoretical cause: naive single-sample approximations in stochastic latent space induce significant variance and bias in the surrogate loss. To address this, we introduce P^3 (Probabilistic Policy Propagation), a distribution-aware optimization framework for VAE-based policies. $P^3$ couples moment-based probabilistic method for stable and efficient learning with sampling-based calibration for robust policy behavior under latent uncertainty. In our experiments, P^3 boosts data efficiency from 64.6% to >96%, reduces convergence steps by >20%. Furthermore, P^3 is evaluated on challenging humanoid parkour tasks and shows an effective foundation for VAE-based PPO. Code is available at https://github.com/ylyem9x/P3_Open.

Contemporary machine learning struggles to learn continually, reuse prior knowledge, and expose a comprehensible internal structure. A recently proposed developmental, gradient-free learning framework addresses these limitations by learning a discrete, topological model of its inputs through local variation and selection, yielding an inherent continual-learning guarantee: new observations refine existing structure without overwriting past knowledge, and without replay buffers or predefined task boundaries. Its extension to visual inputs demonstrated this principle on shape recognition, but relied on a feature representation of limited expressivity that capped recognition accuracy. We introduce a new visual feature representation that encodes shape structure across multiple scales, capturing edge and contour features together with their spatial relations, and integrate it with the network-refinement learning process; we further improve the learning dynamics and the read-out used to predict from the learned model. The study targets two-dimensional shape, with class-incremental MNIST as a controlled, interpretable benchmark in which continual-learning behavior can be measured directly. Our approach substantially increases accuracy over the prior representation, matching or exceeding replay- and regularisation-based baselines at comparable storage while storing no past data, and preserves the framework's defining behavior: earlier-learned classes are retained as new ones are introduced, with no destructive adaptation, and the learned representations remain human-interpretable. What separates the methods is retention: the baselines surrender most of a just-trained class within its own cycle and relearn it afterwards, which ours does not. The significance lies in the manner of learning. The system integrates information one sample at a time while provably preserving its responses to...

Unmanned aerial vehicle (UAV)-satellite cross-view geo-localization matches UAV images against satellite imagery and has achieved impressive accuracy on clean (non-degraded) image benchmarks. In real-world flights, however, UAV observations are frequently affected by adverse weather, illumination changes, platform motion, sensor noise, and compression, while the robustness of existing methods under such degradations remains largely unexamined. In this paper, we present UAVSat-Deg, a large-scale robustness benchmark for degraded UAV-satellite geo-localization, comprising University-1652-Deg and SUES-200-Deg. UAVSat-Deg covers 27 corruption types, including 19 core and 8 compound corruptions, at three severity levels, supports bidirectional drone-to-satellite and satellite-to-drone retrieval as well as multi-height UAV acquisition, and contains more than 11.7 million pre-generated corrupted test images. Benchmarking representative methods under this protocol reveals substantial robustness gaps, particularly under severe and compound corruptions. To address this problem, we propose ReLATE, a Reliable Evidence Learning framework with Adaptive Token Evidence Regulation, which realizes reliability-adaptive feature fusion during descriptor construction. ReLATE estimates a structure-smoothed reliability field over visual tokens, aggregates trustworthy local evidence, and adaptively integrates it into query-derived representations; the regulated query representations are then combined with the CLS-token and GeM-pooled branches to form the final cross-view descriptor. Across both test sets and retrieval directions, ReLATE achieves the best average corrupted-test performance among the compared methods while maintaining competitive accuracy on clean images. The code and dataset will be available at https://github.com/JHC626/ReLATE.

The rapid advancement of video generation models has led to the increasing misuse of image-to-video (I2V) models. Although substantial progress has been made in detecting AI-generated videos, proactive defenses against I2V models remain underexplored. In particular, current proactive defenses against I2V models predominantly rely on gradient-based adversarial attacks, which require defenders to possess GPUs with substantial memory resources (VRAM) to generate adversarial examples. To address this issue, we propose I2VShield, a privacy protection method based on generative adversarial attacks tailored to Diffusion Transformer (DiT)-based I2V models. The proposed method primarily consists of two components: (1) a text-adaptive perturbation generation framework integrating adversarial learning to mitigate computational overhead while maintaining visual imperceptibility; and (2) an untargeted Multimodal Attention Disruption (MAD) attack that exploits the inherent vulnerabilities of DiT-based I2V models, maximizing the deviation of the internal attention features from their clean states. Extensive experiments demonstrate that our approach achieves highly competitive protection performance across various datasets and mainstream DiT-based I2V models, particularly in disrupting spatiotemporal coherence, while substantially reducing computational costs.

Computational AMP discovery is often evaluated through AMP/non-AMP recognition, yet follow-up decisions depend on assay-derived evidence such as target-species potency, hemolysis, toxicity, and selectivity. Existing AMP and peptide benchmarks cover binary recognition, multilabel annotation, assay regression, or broader peptide-model comparison, but they do not jointly place AMP recognition, species-conditioned potency, spectrum, safety-facing proxy endpoints, and cross-endpoint behavior within one sequence-homology-controlled protocol. To address this problem, we introduce AMPBench-MT, a provenance-preserving benchmark that standardizes canonical peptide records and organizes them into binary recognition, species-conditioned pMIC regression, and endpoint-specific potency and safety-facing readouts. Across 161 endpoint-specific model evaluations, high binary performance does not reliably indicate assay-endpoint behavior. Frozen protein-language-model embeddings form the leading pMIC error cluster, while graph and classical regressors remain close. Spectrum labels further reveal that PR-oriented metrics can be misleading under scarce observed negatives, whereas low-toxicity, HC50 hemolysis, and selectivity expose smaller but more assay-facing signals. AMPBench-MT shows that AMP evaluation should move beyond recognition leaderboards toward endpoint-aware evidence auditing. Our proposed benchmark is available at https://huggingface.co/datasets/ZihengZhou06/AMPBench-MT.

Vision-language-action (VLA) models predict sequential actions to execute tasks specified by language instructions, conditioned on visual observations and proprioceptive states. However, how to fuse modalities in VLA models remains an open problem, since robot manipulation involves dynamic phases, such as long-distance movements and close-range interactions, in which the importance of visual observations may vary over time. In this paper, we propose an infer-diagnose-refine (IDR) framework, a model-agnostic framework that can be integrated with diverse VLA architectures for refining action predictions at test time. IDR first infers actions under factual and counterfactual scenarios of visual observations, and then diagnoses the causal effects of visual observations as the estimated dynamic importance, which is finally used to refine the action predictions in a training-free manner. We further design a causality-aware action refiner to realize the IDR framework, including zero-padding interventions for inferring counterfactual actions, norm-based quantification for diagnosing causal effects, and gated residual fusion for refining actions. Extensive experiments on both simulation benchmarks and real-world tasks show improvements in overall performance across multiple VLA backbones, demonstrating the efficacy of dynamically adjusting visual importance at test time.

Fine-grained weight pruning and activation sparsification have emerged as effective approaches for reducing the compute and memory cost of inference for Transformer models. In the moderate-sparsity regime, Gustavson's dataflow provides a natural execution model for exploiting both activation and weight sparsity on vector processors through metadata-driven indexed accumulation. However, existing RVV architectures lack native support for this pattern, forcing kernels to rely on software index decoding and L1-backed indexed memory operations that keep sparse tensor contractions far below their roofline performance bound. We present Ventaglio, a runtime-configurable sparse execution unit coupled with RVV ISA extensions that drives sparse tensor contractions toward their roofline through indexed gather-accumulate-scatter support. Integrated into an open-source vector processing cluster and implemented in 12nm FinFET, Ventaglio accelerates sparse tensor contraction kernels by $6.9\text{--}7.4\times$ over optimized RVV baselines, with only $3.1\%$ area overhead for a cluster of tightly-L1 coupled vector processing elements. We build a performance-accurate instruction-level model of the Ventaglio extension, calibrate it against RTL implementation, and leverage it for scale-out performance analysis on a large $4\times4$ multi-cluster system. Using a DuoGPT-pruned LLaMA-3-8B model with practical $40\text{--}60\%$ dual sparsity, Ventaglio achieves $2.40\text{--}5.25\times$ and $2.06\text{--}3.16\times$ speedup over dense baselines during prefill and autoregressive decoding, respectively.

Pathology foundation models are approaching clinical deployment, yet remain vulnerable to systematic non-biological variation across centres. Differences in tissue preparation, staining and scanning are strongly encoded in their representations, enabling shortcut learning and weakening generalisation across cohorts and institutions. The Robustness Index (RI) quantifies whether local representation geometry is dominated by biology or by non-biological variation, but its count-based formulation discards distance information. We show that adding distance weights changes little because the deeper limitation lies in RI's pooled, fixed-neighbourhood design, which obscures sample-level heterogeneity and effectively evaluates only a model-dependent subset of samples. We introduce the Cross-confounder Robustness Margin (CRoMa), a sample-resolved measure that directly compares distances to cross-confounder biological matches and same-confounder biological distractors. CRoMa recasts robustness as a cohort-wide margin distribution rather than a single pooled score. We evaluated frozen representations from 20 tile-level encoders across three benchmarks and 4 slide-level encoders on a fourth. Rankings by median CRoMa were broadly consistent across datasets, while the underlying distributions revealed substantial within-model heterogeneity. Every tile encoder retained a confounder-dominated lower tail, whose prevalence and severity varied markedly across models. These distinct robustness profiles frame model selection as a Pareto trade-off between typical and lower-tail robustness. Higher CRoMa was also associated with smaller shortcut-induced performance drops after supervised adaptation. By turning representation geometry into a distributional robustness readout that anticipates downstream shortcut susceptibility, CRoMa provides a principled basis for robustness assessment and model selection.

Large language models and multimodal foundation models are enabling medical artificial intelligence (AI) systems to move beyond isolated prediction and undertake multistep clinical tasks that require planning, tool use, memory, iterative correction, and coordination among specialized agents. However, the scope of agentic AI in medicine remains unsettled, and current evaluation practices are not yet aligned with the requirements of clinical use. We conducted a scoping review with systematic evidence mapping across five electronic sources, screened 1,649 exportable records, and provisionally included 557 unique studies that met predefined criteria for goal-directed task execution, tool use, interaction with external resources, feedback-based refinement, or multi-agent collaboration. The included studies describe single agents that use external tools, workflows supported by retrieval and external knowledge, multimodal agents, and multi-agent systems applied to medical question answering, image interpretation, electronic health record analysis, drug safety, and clinical trial prediction. The evidence base remains dominated by public benchmarks, simulated settings, retrospective datasets, and small-scale expert evaluation. Process reliability, evidence traceability, uncertainty, safety, workflow impact, and external validity are evaluated less consistently. Clinical translation will depend on clearer definitions, reproducible evaluation, auditable oversight, interoperable system design, and prospective validation in real-world clinical workflows.

Vision-Language-Action (VLA) models translate natural-language commands into robot action sequences, but leading systems on the LIBERO-Plus robustness benchmark use three- to seven-billion-parameter backbones whose memory demands can exceed embedded robotic budgets. We present CoTinyVLA, a 0.9B-parameter action model on a Qwen3.5-0.8B backbone that obtains that robustness by structuring supervision instead of enlarging the model. Three components target different axes of the problem: dual-view temporal input of 16 history frames per step with textual camera and time markers; hierarchical chain-of-thought (CoT) distillation from a 35B teacher into an episode-level Plan and a chunk-level Think span over task phase, gripper state and next subaction; and paraphrase augmentation expanding 40 base commands into 800 variants. On LIBERO-Plus, spanning 10,030 perturbed tasks across seven perturbation dimensions, CoTinyVLA reaches 90.8% on Spatial, 87.3% on Object, 86.6% on Goal and 80.7% on Long, leading the strongest 7B baseline on all four suites by 4.7, 2.8, 15.9 and 3.0 points, with every margin interval excluding zero. The gains concentrate on the hardest axes of the benchmark: across the eleven published baselines none exceeds 53.2% on Robot Initial States in any suite, whereas CoTinyVLA reaches 73.6% on Goal against 39.9% for the strongest baseline. Ablations show the three components to be separable by perturbation axis, and at a matched image budget how frames are divided between the two cameras and across time accounts for 8.6 points on its own. Closed-loop inference peaks at 2.25 GiB of allocated GPU memory, and paired interventions show the episode Plan to be load-bearing: replacing it with an empty or contradictory span costs 40 to 45 points of success. Structured supervision thus lets a 0.9B backbone exceed all of them. Code: https://github.com/BrainJellyPie/CoTinyVLA

Health AI is evolving from answering questions to agentic systems that converse with patients, reason about health records, and act on their behalf. Primary care guards against diagnostic errors and unsafe care; agents assisting in this domain warrant evaluation against the same risks. Current benchmarks focus on medical knowledge, assessed through isolated question-answering or clinician-facing tasks. PatientAgentBench benchmarks patient-facing agentic healthcare; it evaluates a foundation model, wrapped in an agent with a sandbox of healthcare tools, conversing with a simulated patient. Each conversation is scored by an LLM-as-a-Jury across six dimensions via over a hundred conversation-agnostic, clinician-grounded criteria. To validate alignment, licensed clinicians annotated shared conversations, yielding 79-93% adjacent agreement between jury and expert raters, on par with or exceeding clinician inter-rater agreement. We benchmarked 10 models across four families on the same 1,200 scenarios and found clinical gaps. Triage quality is the most discriminating dimension: pass rates rise from 32% for the weakest models to 88% for the strongest, with agents often acting on administrative requests without clinical screening. Clinical safety and workflow accuracy follow the same pattern: the weakest models fail often, fabricating unexecuted actions, while frontier models fail on only 1-3% of cases, from unverified tool outputs and omitted crisis resources in an emergency. More capable models narrow these gaps but do not close them; the strongest scores only 4.25 of 5 overall. These failures surface only in sustained, tool-using conversations against realistic patient records, confirming that static benchmarks are insufficient as healthcare agentic systems gain autonomy. We release the framework as a reproducible, clinician-validated evaluation standard to help the field close this gap.

Stateful multimodal assistants encode an image once but may answer questions about it many turns later. Attention-guided visual-KV eviction assumes that evidence irrelevant now will remain dispensable, although future questions are unknown. We ask when a visual fact is actually safe to forget and introduce the Causal Visual Memory Audit (CVMA), a paired single-prefill framework that tests what later answers lose when a visual region, the whole image, or prior assistant text becomes unavailable. On VisDial and ConvBench, current attention can rank future-useful regions worse than random even though a diagnostic marginal-utility control shows substantial selection headroom. Aggregate scores hide this failure when later turns do not need vision; controlled and stock-generated histories reveal a second escape route, in which assistant-text KV replaces image KV for facts already stated but not reliably for unstated facts. In the tested stacks, safe forgetting is supported by low future visual dependence or fact-specific verbalization---not by low current attention.

Mechanistic interpretability has largely focused on language models and deterministic toy tasks. Much less is known about how sequence models internally represent latent stochastic dynamics under noisy, partially observed observations. We study this question in a controlled multivariate stochastic volatility setting, where models observe only returns while the ground-truth latent volatility state is known to the researcher. This setting provides a useful benchmark for mechanistic interpretability under partial observability: the latent state is hidden from the model but directly available for evaluation. Across architectures, losses, and output heads, we find evidence for a two-stage computation. Hidden representations encode substantial information about the next latent volatility state, and the output head maps this representation to squared return forecasts. Furthermore, in Transformers, latent-state decodability emerges at identifiable architectural stages whose location depends on the volatility period. In long-cycle regimes, this computation simplifies into an explicit latent-state filter consisting of a learned linear projection followed by $\ell^2$ normalization. Output-head replacement further shows that part of the degradation under noisy MSE training arises from readout misalignment rather than representation failure. These results suggest that stochastic volatility models provide a useful benchmark for mechanistic interpretability under noisy latent dynamics and partial observability.

Multi-agent frameworks built on large language models (LLMs) routinely entangle three logically distinct concerns: who is on the team (organization), how members align (coordination), and which algorithm fuses their work (collaboration protocol). IMACS (Intelligent Multi-Agent Collaboration System) separates the three into orthogonal, independently swappable layers. Classic organizational theory (Belbin roles, Mintzberg coordination, RACI accountability) becomes executable, validated configuration, and the framework places six published collaboration algorithms behind a common interface while exposing roles, coordination, and accountability as independently configurable factors. We use this separation to conduct controlled comparisons in which organizational assignments vary while the collaboration protocol is held fixed. It also turns protocol choice into a variable that can be learned: Adaptive Org Routing, a contextual-bandit meta-protocol, selects a protocol per task under an explicit quality-cost tradeoff, outperforms every fixed protocol in a controlled study, and trains online on real benchmark and LLM-judge rewards. The ablations expose a mechanism. Accountability placement changes outcomes exactly when the protocol routes the deliverable through the accountable agent, and the winning placement flips across model families, so organizational design cannot be hard-coded; it must be revalidated, or learned, for each model binding.

Covariate shift across training-data partitions biases model selection and parameter estimation in cross-validation, lifelong learning, and federated learning. We propose \textit{Partition-Induced Covariate-shift Correction} (\texttt{PIcsC}), a Fisher information-based regularization framework that mitigates distribution mismatch between data partitions and a reference distribution. \texttt{PIcsC} approximates partition divergence using the Fisher Information Matrix (FIM) and incorporates the resulting statistic as a regularizer during optimization. The same formulation applies to both centrally partitioned datasets (batches or cross-validation folds) and inherently distributed data (federated clients or decentralized nodes), requiring only partition-local gradient statistics rather than raw data. We further introduce a conditional adaptation mechanism that combines FIM shift with KL divergence to detect significant distribution shifts and activates regularization only when necessary. Experiments on more than 40 datasets demonstrate consistent improvements under both natural and synthetic covariate shift. On fragmented batch and fold settings, \texttt{PIcsC} reduces fragmentation-induced performance degradation by more than 20\% and 25\%, respectively. On seven federated learning benchmarks, it consistently outperforms FedAvg, FedProx, and SCAFFOLD by 3 -5 percentage points without requiring client-specific personalization. These results demonstrate that Fisher information provides an effective and unified mechanism for mitigating partition-induced covariate shift across both centralized and distributed learning.

Long-horizon conversion prediction under delayed feedback creates a two-clock, two-window learning problem in online advertising. A short base observation window releases recent clicks on the click clock before their outcomes mature, whereas conversions continue to arrive on the conversion clock throughout a longer target conversion window. The click clock provides timely but partially observed status supervision. The conversion clock reveals long-tail delays, but the delay composition within an arrival-time slice is weighted by historical click cohorts with different traffic volumes and target-window conversion rates. We present TWICE, a framework that factorizes long-horizon post-click conversion rate (CVR) into a target-window conversion probability and a grouped elapsed-delay cumulative distribution function (CDF). The two clocks provide complementary supervision. Click-clock records train the target-window CVR head through a current-status likelihood over the base observation window. Newly arrived conversions train the delay model on the conversion clock. To account for the cohort mixture, TWICE uses fixed click-time predicted CVR (pCVR) mass as cohort exposure in an arrival-conditioned likelihood. This accounts for differences in cohort traffic and conversion propensity. The resulting aggregate records are self-contained. A single learned CDF produces monotone predictions for all requested horizons up to the target conversion window. Serving requires neither historical lookup nor convolution. Experiments on a public benchmark and an industrial advertising dataset demonstrate the effectiveness of TWICE. In an online A/B test in Kwai's advertising system, TWICE increased expected revenue, revenue, and conversions by 2.486%, 1.858%, and 2.061%, respectively. It was subsequently deployed to full traffic.

Language-model agents are increasingly deployed under standing instructions: a system prompt, a policy file, or a skills document is placed in context, and the agent is trusted to let it govern every action that follows. Existing benchmarks rarely test this deployment pattern directly; they measure whether an agent can complete a task, not whether a long, binding policy document actually constrains its behavior over an extended tool-use horizon. We present HANDBOOK.md, a benchmark of 65 agentic tasks modeled on how enterprise employees follow company handbooks. Each task places an agent in a self-contained company environment, a file workspace together with mock email, chat, calendar, issue-tracking, and commerce services exposed over the Model Context Protocol, and instructs it to carry out routine professional work governed by an expert-written standard operating procedure of 20 to 124 pages. Tasks span five domains (finance, medical billing, insurance, logistics, and HR) and ten fictional companies. To resist memorization, every task modifies one of ten base handbooks, altering the specific rules and thresholds on which grading turns, so no two tasks share a policy. Grading is fully deterministic: each task carries a rubric of programmatic criteria (824 in total) that check both that required actions occurred and that prohibited actions did not. Under strict grading, where a trial passes only if every criterion is satisfied, the best of thirty evaluated model configurations passes 36.2% of trials, and most frontier configurations remain below 25%. Failures follow consistent patterns: agents let a plausible in-environment request override the standing policy, perform a required check and then act against its result, lose rule details over long horizons, and report compliance they did not achieve. We release all tasks, environments, and the evaluation harness.

Successfully automating dexterous, long-horizon robotic manipulation requires frameworks capable of both high-level reasoning and fine-grained execution. Traditional task and motion planning (TAMP), while excellent at symbolic planning, is often brittle in contact-rich operations. Simultaneously, imitation learning (IL), while effective in manipulation tasks with visual feedback, is limited by its low capability in spatial generalization and multi-stage operation. To reconcile their complementary strengths and limitations, we propose DR-LfD (Decomposed and Reorganized Skills Learned from Demonstrations), a framework that seamlessly integrates visuomotor policies into a TAMP-gated decision-making system. Based on contact relationships, DR-LfD decomposes human demonstrations into atomic skills, which are reproduced as visuomotor policies or object-centric primitives. The initiation, termination, and constraints of the visuomotor policies are carefully modeled and implemented in a TAMP-compatible form, enabling reorganization of skills learned from different sources. DR-LfD transforms the learning problem from one requiring exponential demonstration data over possible skill sequences to one whose demonstration burden scales with the number of distinct skill types, with limited data for each skill. Through comprehensive real-world and simulation benchmarking across diverse scenarios, we demonstrate the strong performance of DR-LfD on tasks involving multiple steps, unseen setups, and physical constraints. Project website: https://dr-lfd.github.io/DR-LfD-website.

We introduce RDVSv2, a large-scale benchmark for RGB-D video salient object detection (RGB-D VSOD) with dense frame-level annotations. Existing datasets in this emerging field are often limited in scale and annotation quality, while also relying on less geometry-consistent depth cues. To address these limitations, RDVSv2 is built from publicly accessible stereoscopic online videos and contains 249 video sequences with 29,077 annotated frames. It includes depth maps derived from stereoscopic videos, together with frame-wise salient object masks annotated with eye-tracking guidance. Compared with existing datasets, RDVSv2 is much larger in scale and covers more diverse and challenging scenarios. In addition, we establish a strong baseline for RGB-D VSOD based on Segment Anything Model 2 (SAM2). Specifically, we employ a parameter-efficient fine-tuning (PEFT) strategy to adapt the SAM2 encoder to jointly encode RGB, depth, and optical flow cues. Extensive experiments show that RDVSv2 is substantially more challenging for existing RGB-D VSOD methods. Meanwhile, the proposed baseline achieves state-of-the-art results on RDVSv2 and existing RGB-D VSOD benchmarks. We hope that RDVSv2 and the provided baseline will serve as useful resources for future research on RGB-D VSOD and related multi-modal video understanding tasks. Our dataset and code will be available at https://github.com/ltynick/RDVSv2.

Autonomous multi-vehicle racing requires real-time planning of diverse competitive behaviors in intense interactions. Existing planners often struggle to balance strategic diversity and computational efficiency. To address this challenge, we propose Sampling-based Game-Theoretic Planning (SGTP), a real-time framework that combines game-theoretic reasoning with GPU-accelerated sampling of control sequences and dynamics rollouts. Sampled trajectories are ranked using a game-aware cost to capture competitive interactions and generate diverse racing behaviors. Our planner then performs feasibility selection by explicitly enforcing track-boundary and dynamic collision-avoidance constraints, ensuring safe and reliable transitions between racing strategies. Extensive simulations on challenging tracks show that SGTP achieves a 95.24% win rate and a 99.35% task-completion ratio in highly interactive races, with a mean computational time of 0.095 s over multiple iterative solving steps. We also demonstrate the successful application of SGTP in large-scale scenarios with up to 10 agents. We release our code and provide an open-source benchmark of multi-agent autonomous racing algorithms to facilitate future research. Project page: https://sgtp-racing.github.io/.

Learned restriction maps in sheaf graph neural networks are often treated as proof that the model has discovered useful edge geometry. That conclusion does not follow from parameter movement or from a post-hoc ablation: both can show how one checkpoint is organized while leaving open whether learned transport still helps after the rest of the model adapts. We separate these claims with two estimands. Checkpoint reliance intervenes on the maps of a fixed predictor; protocol-relative replacement retrains matched families that remove map capacity, edge variation, or persistent edge assignment. A task-null theorem shows why the claims can diverge: labels identify only the transported classifier directions, leaving $d^2-d$ invisible degrees of freedom in every full $d\times d$ map. An exact frame model then gives the boundary at which reliance becomes unreplaced task value. Label-only training realizes the predicted separation, while audits of public NSD, DNSD, and Directed Sheaf Neural Network (DSNN) implementations recover both replaceable and unreplaced transport regimes on real graphs. All five DNSD benchmarks exhibit fixed-checkpoint reliance. After retraining, assignment-breaking or shared-map controls recover Full performance on four; Roman-Empire retains a $.0675$ advantage over continually resampled assignment and a $.0391$ advantage over a parameter-matched shared map across ten official splits. Thus, a learned map can govern a fitted computation without constituting indispensable edge geometry. Claims of learned transport should pair checkpoint interventions with matched retraining.

Memory has evolved into a foundational architectural dimension in large language models (LLMs), shifting from an implicit byproduct of computation to a spectrum of explicit, controllable mechanisms. While recent advances introduce diverse strategies---spanning transient attention, recurrent state dynamics, parameter-efficient adaptations, and scalable lookup storage---this rapid evolution has led to a highly fragmented research landscape. In this survey, we present a systematic, architecture-centric taxonomy of memory in LLMs. Our framework characterizes memory along three orthogonal axes: representation (implicit versus explicit), update dynamics (offline versus online), and persistence (short-term versus long-term). We further formalize the granular mechanisms dictating memory writing, routing, state transitions, and consolidation. This unified perspective elucidates the conceptual boundaries between computation-coupled and independently addressable memory, effectively bridging disparate architectural paradigms. Additionally, we critically analyze hybrid memory architectures, system-level efficiency trade-offs, and multi-dimensional evaluation methodologies. By consolidating these scattered advancements into a cohesive framework, this survey charts the trajectory of memory-centric LLM design and provides a principled foundation for future innovations in scalable and adaptive language modeling.

Medical image visualization requires volumetric rendering algorithms that preserve anatomical fidelity while maintaining high rendering speeds. To address the high computational cost of large volumetric datasets, we propose a Gaussian-based volumetric representation for efficient visualization of dense medical volumes without compromising structural and radiometric details. We optimize the proposed representation using Monte Carlo volumetric estimation, which enables training on a highly sparse subset of voxels while maintaining consistency with the dense volumetric objective. In addition, we introduce a curriculum learning strategy that progressively incorporates structured slice-based sampling during training. Sparse voxel samples provide an early global coverage of the volume, while slice samples capture spatially correlated regions that aid geometric structure and texture continuity. This combination enables the Gaussian representation to learn anatomical details of various structures and corresponding textures from sparse supervision while significantly reducing the computational cost associated with dense voxel processing. The learned representation supports slice-based rendering methods such as shear-warp volume rendering, enabling efficient visualization of multimodal medical datasets including MRI and Cryosection volumes while preserving anatomical structures. Using sparse supervision, our method achieves up to 43.86 FPS rendering with a compression ratio of 11.31:1.

India is a vast nation of over 1.4 billion people, varied by hundreds of diverse and locally specific traditions and cultures and 22 officially recognized languages. Large language models (LLMs) are now being deployed on a massive scale throughout the mainland as well as in remote villages. However, the common benchmarks - MMLU, BIG-Bench, and TruthfulQA are almost exclusively English- and Western-centric. They do not identify those safety, fairness, and accuracy failures unique to the Indian context. That is the gap Inspect India Evals seeks to fill. It is an open-source framework built on top of UK AISI's Inspect AI platform. It has six benchmarks: Multilingual MMLU across sixteen Indian languages, BharatBBQ (our adaptation of BBQ for Indian social bias), a safety evaluation for Digital Public Infrastructure, a multilingual safety test using harmful prompts in Indian languages, a multi-turn jailbreak resistance test, and an Indian cultural knowledge benchmark scored using LLM-as-judge rubrics. In this study, we tested five open-weight models ranging from 8B to 32B parameters. Sarvam-M 24B and Gemma 2 27B came out on top, both scoring 80% on the composite India Fairness Index, with Sarvam-M even beating larger 32B models on Indian cultural knowledge and DPI safety compliance. All models scored 100% refusal on Multilingual Safety, whereas DPI safety varied from 20% to 100%. The framework is public. It's built to work with the UK AISI registry. Anyone can reproduce or extend this work.

Reflections from shiny objects and glass facades naturally extend the field of view of a camera, capturing the surrounding environment without the need to pan the camera or acquire a full panorama. We propose PanoLess, a Gaussian-splat-based framework that reconstructs the surrounding environment as a distant illumination map from images captured on only one side of a reflective surface. PanoLess leverages surface-aligned 2D Gaussian splats with deferred shading to recover accurate per-pixel normals and reflection cues, which are fused into a neural cubemap representation of the environment. In addition, PanoLess produces a visibility map that explicitly denotes which regions of the environment are supported by the partial reflective observations. Unlike existing inverse-rendering and reflection-aware Gaussian-splatting approaches, which typically require full 360-degree coverage and struggle under incomplete views, PanoLess enables consistent, physically grounded illumination estimation from partial-view input. We show that PanoLess achieves high-fidelity and geometrically consistent environment reconstruction, outperforming reflection-aware baselines on a new custom synthetic benchmark and publicly available datasets, and demonstrating generalization to real-world reflective captures.

Long-context recall in linear-time sequence models highlights a tradeoff in how they write to memory. State-based linear models, such as state-space models (SSMs) and linear Transformers, write densely, updating the entire state for each newly arrived token, which leads to interference and makes specific past tokens hard to recover. Sliding-window attention (SWA) exhibits the opposite behavior: it writes sparsely by storing explicit token representations, but only within a fixed window, so recall drops once the relevant token is evicted. Interpolating between these models, we introduce Raven, a linear-time sequence model that maintains a fixed set of memory slots and, at each step, decays and updates only a selected subset via learned, input-dependent routing. This lets Raven mitigate SWA's position-based overwriting and hard eviction while reducing interference from dense state updates in SSMs, thereby preserving long-range content much more effectively. Across recall-intensive benchmarks, Raven is competitive with or outperforms prior linear-time baselines, achieving strong long-context recall where both SWA and SSMs sharply degrade. It remains effective when extrapolating to context lengths as large as 16x its training length, with similar gains in hybrid architectures.

Data quality profiling -- computing missing-value rates, duplicate fractions, outlier densities, and functional-dependency violations -- is foundational for data-centric AI pipelines, yet exhaustive scans over millions of rows are prohibitively slow for near-real-time monitoring. Progressive sampling is the standard alternative; the open question is which strategy best preserves profile fidelity at scale. We benchmark nine sampling strategies -- blind (random uniform, geometric, Yamane, cluster) and proxy-guided (Metropolis-Hastings, DAG, stratified by column type or quality score, importance-weighted) -- on three real-world datasets (NYC 311, NYPD arrests, UCI Adult; up to 500K rows), an IoT sensor stream (2.3M rows), two ultra-large real datasets including Ultra-Marathon Running (up to 7.4M rows), and synthetic data scaled to 5x10^6 rows. Contrary to the assumption sharpens estimates, blind representative samplers dominate uniformly. At a 5% budget, random uniform achieves 0.49% mean relative error on NYC 311; DAG-guided MCMC yields 19.5% (approx. 40x worse), and across all real datasets DAG is 11-49x worse (Wilcoxon W=0, p=0.002, n=9 pairs). Cluster sampling matches random uniform (MRE 0.110 vs. 0.111); proxy-guided methods share DAG's failure mode (MRE 0.20-0.35). At scale, random uniform is near-linear (O(N^{0.964})) while DAG is super-linear (O(N^{1.272})), running 28--47x slower on ultra-large data with 6x worse accuracy. The root cause is an IQR proxy mismatch: proxy-guided samplers over-pursue numeric outliers, while quality defects concentrate in categorical columns invisible to the proxy. The actionable finding: representativeness, not domain knowledge, determines sampler quality -- schema-free random uniform or cluster sampling suffices for production-grade quality profiling at scale.

The same episode of atrial fibrillation is a minor finding in a healthy adult and grounds for anticoagulation in an elderly patient with hypertension: identical signal, opposite decision. Naming the rhythm is only the start; what determines a patient's outcome is the judgement that follows -- what the arrhythmia is across the whole record, what it means for this patient, and what should be done about it. Recent work pairing large language models with the ECG stops short of this, reading one recording without assembling a patient-level finding; and agentic systems built around it either receive the arrhythmia a device has already detected or target a different diagnostic task, stopping before the decision this task requires. We formulate patient-level arrhythmia decision support as a task and present Cardiologent, a multi-agent system that spans it from detection to decision. An agent for each signal -- a single ECG lead and the photoplethysmogram a wearable acquires -- grounds its window reading in measured features rather than a bare label; the readings are assembled into the patient's rhythm profile and, with the patient's own data, reasoned against clinical guidelines retrieved for the case, with a critic checking each conclusion against the guideline it cites. We evaluate the clinical decision rather than the report, across integrated diagnosis, clinical significance, and urgency and management. Cardiologent scores highest on every axis, first on every patient-level task under both cardiologists and an at-scale LLM judge -- whose agreement with the cardiologists (ICC 0.74, 0.66) matches theirs with each other (0.67). Because each conclusion traces to a cited guideline and is validated against expert cardiologists, it yields decisions a clinician can audit rather than act on blindly -- a step toward use in continuous monitoring.

Joint-Embedding Predictive Architectures (JEPAs) learn world models by predicting in representation space rather than reconstructing pixels, making them a natural backbone for latent model predictive control from offline demonstration logs. JEPA-style training optimizes short-horizon latent prediction, whereas planning requires a multi-step ranking of imagined futures by goal progress. Prior JEPA planners often inherit that ranking from embedding geometry, typically latent Euclidean distance, which arises as a byproduct of representation learning rather than as a progress cost mined from the logs. We propose temporal-distance JEPA (TD-JEPA), which retains the LeWM encoder--predictor backbone and mines a directed temporal cost from reward-free trajectories: same-trajectory step order supplies positive targets, cross-trajectory pairs act as heuristic negatives, and a rollout-consistency term matches the planner horizon. The mined supervision serves two roles: as the deployed planning cost when progress is topological, and as a representation signal that improves Euclidean planning when contact geometry dominates. Under locked evaluation, deploying the mined cost raises Two-Room success to 100.0% versus LeWM's 97.4%, while shared Euclidean planning on the same temporally trained checkpoint raises OGB-Cube by 14.2 points over LeWM and improves Push-T. Against LeWM and the concurrent RC-aux baseline under locked evaluation, TD-JEPA matches or exceeds both methods on every environment. Ablations show that the directed head, cross-trajectory negatives, and rollout consistency each contribute. TD-JEPA narrows the train--plan gap for JEPA world-model planners by discovering temporal progress structure in offline logs and co-designing cost form with plan-time deployment. Code is available at https://github.com/HKBU-KnowComp/TD-JEPA.

Dataset distillation aims to synthesize compact datasets that can approximate the performance of full-data training while significantly reducing computational and storage costs. However, diffusion-based distillation methods often struggle to preserve structural coherence and generalization, especially in visually complex domains. This issue often stems from latent prototypes that are weakly aligned with class-discriminative regions and contaminated by irrelevant background, thereby degrading generation quality and generalization. To address this limitation, we propose a saliency-driven distillation framework that constructs class-discriminative latent prototypes to enhance representativeness and generalization. The framework proceeds in two stages: (1) ensemble Grad-CAM saliency is used to construct prototypes emphasizing high-confidence regions, and (2) hard prototype refinement is then applied to construct challenging yet class-consistent prototypes, thereby enhancing discriminability and diversity. Importantly, the diffusion backbones (e.g., LDM and DiT) remain frozen; only lightweight classifiers used for saliency extraction are trained. Extensive experiments across multiple benchmarks demonstrate consistent performance improvements over strong baselines. Code will be released.

Collective intent prediction in multi-agent systems focuses on predicting the shared objectives and future behaviours of groups of interacting agents. The problem is particularly challenging because collective intent emerges from complex interactions, evolving cooperation structures, and long-term behavioural dependencies among heterogeneous agents operating in dynamic and partially observable environments. Furthermore, functional roles adopted by agents are often latent, may change over time, and are rarely available as explicit annotations, making the learning of coordinated group behaviours significantly more difficult. To address these challenges, this paper proposes a Meta-Role Temporal Graph Network (MR-TGN) framework for collective intent prediction in multi-agent systems. The proposed framework models agents as dynamically evolving graph entities and employs temporal memory mechanisms to encode historical interactions and coordination behaviours. To capture higher-level behavioural knowledge, MR-TGN introduces a memory-enhanced meta-role learning mechanism that derives latent role representations from agent-centric behavioral representations without requiring explicit role labels. An evaluation methodology for early collective intent prediction is proposed to assess prediction accuracy and timeliness, enabling realistic evaluation of the model's ability to anticipate collective objectives during the early stages of mission execution. Experimental results on representative multi-agent scenarios demonstrate that the proposed framework consistently outperforms competitive baselines and achieves effective early prediction of collective intents in dynamic and adversarial environments.

Sensory advertising evokes human senses through visual cues, enabling audiences to mentally simulate experiences and increasing persuasive impact. Despite the recent increase in using AI in generating and understanding creative and persuasive content, how advertisements visually evoke sensations remains largely unexplored. In this work, we introduce the first study of understanding, evaluating, and generating sensory ads. We introduce the Sensory Ad dataset, and define sensation classification tasks (SenseClass) to benchmark LLMs and MLLMs. We further propose SenseScore, an automated evaluation metric for sensation evocation achieving strong agreement with human judgments. Finally, we introduce the Sensory Ad Generation (SenseGen) task and propose SAGA, a multi-agent framework that improves message image alignment, sensory evocation, and persuasion. Our work establishes a foundation for sensory-aware visual persuasion.

Hyperspectral imaging (HSI) is useful for material discrimination, but operational mine screening also depends on how many false alarms must be inspected before targets are found. This paper studies PFM-1 landmine detection in unmanned aerial vehicle (UAV) visible and near-infrared (VNIR) HSI using spectral angle mapper (SAM), matched filter (MF), adaptive coherence estimator (ACE), and constrained energy minimization (CEM). We compare a ground-measured SVC signature, a fully informed in-scene core-pixel signature, and a simulated human-in-the-loop signature bootstrap. Besides receiver operating characteristic area under the curve and average precision, we report target-discovery curves and spatial candidate-review counts. Full-review bootstrapping reaches the fully informed in-scene signature case after all seven target regions are verified, but the required inspection effort varies strongly: ACE confirms all regions in two rounds and nine candidate inspections, whereas the SAM variants need thousands of candidate reviews for their final target locations. Code is available at https://github.com/SagarLekhak/IEEE_WHISPERS_2026_UAV_HSI_PFM1.

Training large language models (LLMs) to act in long-horizon games is a promising step toward generalist decision-making, yet reinforcement learning with verifiable rewards (RLVR) relies on sparse final rewards that reveal little about which decisions determine success. Denser process signals could supply this missing turn-level credit, but existing sources are hard to keep both cheap and accurate. We observe that changes in a game solver's state value reveal whether an action advances the state toward success. Building on this insight, we propose CAST (Credit Assignment from Solver Teachers), which converts these value changes into solver advantages and injects them into RLVR as turn-level signals. We further show that, under a soft-optimal solver assumption, maximizing the solver advantage is equivalent to on-policy distillation from the solver, requiring only scalar values rather than teacher logits. Across Sokoban, Minesweeper, and Rush Hour, CAST outperforms all trained baselines on every game under both in-domain and unseen-difficulty evaluation and achieves the highest average zero-shot performance on ALFWorld and WebShop. Our code is available at https://github.com/Wloner0809/CAST.

Video editing is fundamentally message-driven: even from the same source footage, the selected shots change depending on the narrative the editor wishes to convey. Benchmarks for a closely related task, video summarization, reduce editorial intent to a single, message-agnostic notion of saliency and thus do not account for this diversity. For evaluating message-driven video editing, we present \textbf{MEDit-Bench}, a dataset and benchmark, which pairs long-form videos with multiple editing messages and multiple professionally produced edits per message, demonstrating that different messages yield substantially different edits from the same source. We define an automatic evaluation protocol based on temporal alignment metrics, and find that an LLM-as-a-judge preference, a natural proxy for narrative quality, is unreliable for this task due to severe position bias. We additionally annotate each message with ambiguity and contextfulness scores, and show that both dimensions negatively correlate with model performance, establishing message difficulty as a meaningful stratification factor. Experiments with state-of-the-art MLLMs and reinforcement fine-tuned baselines show that while strong models approach human temporal alignment at lenient thresholds, all models fall behind humans at stricter criteria. A human perceptual study further confirms a large quality gap, with professional human edits remaining consistently preferred over model outputs.

Optimization over the Stiefel manifold plays a significant role in various machine learning tasks. Existing methods either use the retraction operators, requiring costly orthonormalization for large-scale matrices, or employ landing methods that rely on careful step size selection and penalty parameter tuning. To address these challenges, we propose a retraction-free and penalty parameter-free algorithm that directly lands on the manifold. By leveraging the strongly-convex-like property of the quadratic penalty function and the proximal smoothness of the Stiefel manifold, we establish global convergence guarantees with the best-known iteration complexities under both constant and diminishing step sizes. Then, we reformulate the low-rank adaptation (LoRA) fine-tuning problem for large language models as a manifold optimization problem, introducing Manifold-LoRA for geometry-accelerated adaptation. This approach employs the proposed landing technique and a carefully designed step size strategy to accelerate the training process. Numerical experiments on benchmark datasets demonstrate the efficiency and strong downstream performance of the proposed method.

Tensorial multi-view clustering (TMC) has achieved strong performance due to its ability to capture high-order correlations across multiple views. Most existing t-SVD-based TMC frameworks apply the Fast Fourier Transform (FFT) along the sample mode to impose frequency-domain low-rank constraints. However, we reveal that this widely adopted design critically relies on an implicit ``periodicity assumption'' induced by the sample arrangement. When samples are ordered by class, neighboring indices tend to be semantically similar, creating artificial local continuity along the sample mode and a favorable spectral structure for FFT-based low-rank regularization. Once this ordering is removed by random permutation, existing t-SVD-based TMC methods suffer severe performance degradation. This strong sensitivity to class ordering conflicts with the permutation-invariant nature of clustering and indicates that part of the reported performance may be attributed to a privileged sample arrangement rather than genuine high-order structure modeling. In this paper, we systematically investigate this phenomenon and its underlying algebraic and spectral mechanisms. To address this fundamental flaw, we further propose a graph-spectral low-rank tensor learning framework based on the Graph Fourier Transform (GFT), which replaces the fixed Fourier basis along the sample mode with a data-driven graph spectral basis, thereby capturing the intrinsic manifold structure without relying on a particular sample ordering. Moreover, we develop an anchor-based variant to address large-scale datasets efficiently. Extensive experiments on various benchmarks validate our findings and demonstrate the competitive or superior performance of the proposed methods compared with state-of-the-art TMC approaches.

Real-world tasks often require models to learn from task-specific context rather than relying only on pre-trained knowledge. While recent work has highlighted this capability as context learning, existing evaluations mainly focus on textual contexts. In many practical settings, however, the context to be learned from is multimodal: scientific findings are conveyed through figures and tables, financial indicators are scattered across converted reports, and spatial decisions depend on maps, scenes, or web pages. We introduce CLBench-V, a benchmark for multimodal context learning that addresses the difficulty of localizing where context use breaks down by organizing tasks around three dimensions: context grounding, new information application, and new knowledge learning. CLBench-V combines converted public benchmarks with newly constructed datasets spanning domains such as science, finance, long-document understanding, spatial reasoning, and web-based visual question answering. To reduce the cost of constructing domain-specific context-learning tasks, we further use automated construction and filtering procedures for our newly built datasets. Across 3,443 instances and six recent multimodal models, the best overall score is only 0.2847, indicating that multimodal context learning remains far from saturated. Moreover, InternVL3.5-30B-A3B performs best on context grounding and new knowledge learning, while Qwen3.5-Plus performs best on new information application. We further analyze judge reliability, context length, image count, and representative failure cases. Code is available at https://github.com/IamLihua/CLBench-V.

The quadratic cost of self-attention makes long-context inference prohibitively expensive, and proxy-based block-sparse attention has become a practical remedy. Existing methods typically rely on a proxy to predict a binary sparse mask and a kernel to consume this mask and perform sparse attention computation. Such an approach is effective under moderate budgets. However, as the budget tightens, the estimated proxy inevitably drops some salient blocks, while the kernel can only apply the sparse mask mechanically, leading to an evident drop in model accuracy. We propose CoSA, a two-stage training-free Sparse Attention under proxy-kernel CO-design, which couples a Kernel-Aware Proxy (KAP) with an Ordered-Skipping Kernel (OSK). In the first stage, the KAP selects blocks under a moderate budget and produces an ordered mask that prescribes the order in which KV pages are visited in the kernel inner loop. In the second stage, the OSK applies this mask and skips more blocks under a tightened budget given online-softmax statistics. Across mainstream LLM backbones and long-context benchmarks, CoSA attains higher accuracy at lower budgets. Impressively, CoSA achieves a 4.93$\times$ attention speedup and reduces end-to-end Time-to-First-Token by 2.53$\times$ under a context length of 128K with negligible performance degradation.

Single-cell ribonucleic acid sequencing (scRNA-seq) is a foundational technology for precision-medicine workflows that contribute to United Nations Sustainable Development Goal 3 on Good Health and Well-being, and unsupervised clustering is the analytical step that turns raw expression matrices into interpretable cell populations. Practitioners therefore face a recurring engineering decision: is an additional deep representation stage worth its compute and tuning cost, or do classical principal component analysis (PCA) pipelines already suffice? We address this question with a diagnostic benchmark of nine clustering pipelines on ten real datasets (90-5,685 cells, 19,046-41,480 genes, 4-11 cell types), augmented by a partial scVI V2 specialized comparison on seven datasets. The protocol integrates Optuna hyperparameter search, repeated-run robustness, Friedman/Wilcoxon-Holm/TOST testing, and Sobol total-order sensitivity analysis. The contrastive autoencoder achieved the highest mean Adjusted Rand Index (0.7872), but Holm-corrected tests did not establish dominance over the strongest baselines. Per-dataset analysis reveals three reproducible regimes: probabilistic variational autoencoder (VAE) variants help on the smallest datasets, deep autoencoders win on mid-scale data with multi-batch or many-type structure, and classical PCA pipelines remain competitive when linear projection already captures the dominant variation. Sobol indices identify learning rate ($S_T=0.70$) and latent dimensionality ($S_T=0.56$) as the dominant variance contributors, indicating where limited tuning budgets should be allocated. The contribution is therefore a dataset-aware and compute-conscious decision framework for biomedical AI pipelines supporting sustainable healthcare analytics, rather than a universal superiority claim.

We present FunnelAL, a retrieve-then-rank active learning system for single-class discovery, which adapts the multi-stage funnel architecture of industrial recommender systems to data annotation. Large-scale supervised learning faces two challenges: efficiently finding relevant samples in a massive corpus, and distinguishing true positives from visually confusable negatives when embeddings do not cleanly separate classes. Conventional active learning offers a principled framework for reducing annotation cost, yet it treats sample selection as a single-stage process that addresses neither challenge efficiently. FunnelAL decomposes the problem into cascaded stages. Starting from a single positive and negative example, the system iterates through: (1) embedding-based retrieval scoring that narrows the corpus to a manageable candidate set; (2) a precision-triggered ranking stage that exploits a learned ranker (RankNet) while batch precision remains high, then automatically blends in committee-based exploration (QBC) once returns diminish; and (3) feedback from the annotator's labels that refines both stages in subsequent iterations. We evaluate on three diverse image classification benchmarks. With a perfect annotator, FunnelAL attains the best final F1 on all three benchmarks, the best annotation efficiency (first in AULC), and the fewest annotation rounds. The most recent single-class discovery methods (GAL, PF-MA) at best match its final quality, and only at consistently higher labeling cost. Under annotator labeling errors at realistic rates, FunnelAL remains first or statistically tied for first while classical uncertainty-based methods degrade two to three times faster. Our work provides a concrete bridge between multi-stage recommender systems and active learning.

Conformal prediction (CP) provides distribution-free uncertainty quantification, and its extension to graphs is an active research direction. Diffused Adaptive Prediction Sets (DAPS) is a widely used graph-aware diffusion baseline, propagating Adaptive Prediction Sets (APS) non-conformity scores along edges with a uniform coefficient $λ$. We identify a fundamental shortcoming of this design: the uniform low-pass diffusion presupposes graph homophily and proves detrimental on heterophilic graphs, enlarging the mean prediction-set size by up to 10.6% relative to plain APS. To mitigate this, we propose HeAD-CP, a family of node-wise diffusion variants whose coefficients are determined by a label-free local-homophily estimate derived from the GNN softmax. Three variants, namely signed-$γ$, edge-compatibility, and a DAPS-baseline-with-correction, are most effective at extreme heterophily, intermediate heterophily, and moderate-to-high homophily, respectively, and all preserve the marginal coverage guarantee. On ten benchmarks, the HeAD-CP family stays at or below plain APS on every dataset, while DAPS exceeds APS on six. The post-hoc oracle over the family improves over DAPS on 8/10 datasets at $p<0.01$ (paired Wilcoxon), with the largest gains on heterophilic graphs (10.3% on Texas); on the two homophilic datasets where DAPS still wins (CiteSeer, PubMed), it retains a marginal advantage of at most 0.002, statistically insignificant on CiteSeer ($p=0.23$). Designing a calibrated label-free selector that approaches this oracle is the main outstanding empirical question.

Item Response Theory (IRT) has recently been proposed as a framework for evaluating large language model (LLM) benchmarks by separating a model's latent ability from the properties of individual benchmark items. Existing neural IRT approaches, including PSN-IRT, estimate these quantities using point estimates, limiting uncertainty quantification and downstream statistical inference. We introduce Laplace-PSN-IRT, a post-hoc last-layer Laplace approximation that augments a trained PSN-IRT model with approximate Bayesian posterior inference, recovering calibrated uncertainty over model ability and item difficulty without retraining. The resulting posterior enables credible intervals, probabilistic comparisons between models, and propagation of parameter uncertainty into Fisher-information-based item selection. We show that most pairwise comparisons among 12 models on a standard LLM benchmark leaderboard are not statistically distinguishable despite differing point-estimate ranks. We further show that point-estimate Fisher information can become nearly zero for many benchmark items because it is evaluated at a single reference ability, whereas posterior-expected Fisher information remains substantially more stable across the ability range. Finally, posterior-expected Fisher information more accurately recovers full-benchmark ability rankings from small benchmark subsets in most experimental settings while matching point-estimate performance for the smallest subsets. We validate the calibration of the approximate posterior using held-out predictive coverage and find that modeling item difficulty as random while treating item discrimination as fixed produces well-calibrated uncertainty in this architecture.

Medical world models offer a framework for extending medical artificial intelligence beyond static prediction by representing evolving patient states and modelling how they change over time and in response to clinical interventions. This Review defines the conceptual boundaries, technical foundations, application domains, and evidence requirements of the field through a structured narrative synthesis with reproducible evidence mapping.We screened 1,455 unique records and assembled a corpus of 98 sources, including 14 studies that met a strict empirical definition of a medical world model. The field is organised around four capabilities: patient state representation, temporal dynamics modelling, intervention-conditioned simulation, and clinician-supervised planning. Evidence spans medical imaging, longitudinal electronic health records, treatment response modelling, physiological and multimodal state modelling, ultrasound and surgical interaction, and population and health-system simulation; clinical digital twins are treated as a cross-cutting integration framework.Current studies provide early evidence of technical feasibility for trajectory forecasting and comparison of candidate interventions, but most remain retrospective, task-specific, or preclinical. The evidence base is further limited by incomplete longitudinal intervention data, inconsistent action semantics, limited causal identifiability, long-horizon error accumulation, inadequate uncertainty estimation, and limited external validation. Clinical translation will therefore depend on precise intervention representations, robust causal and mechanistic grounding, calibrated trajectory-level uncertainty, safety-constrained planning, and prospective multicentre validation against clinically meaningful endpoints.

Standard offline reinforcement learning (RL) algorithms typically assume that the actions in the dataset are observed without error. However, in many real-world applications, the true actions are unobserved and only noisy proxies are available, causing existing RL methods to yield biased and potentially misleading conclusions. We study off-policy evaluation in infinite-horizon discounted Markov decision processes with hidden actions. By leveraging the next-state variable as a natural proxy for the unobserved action, we establish identification of the policy value and propose an influence-function-based estimator called LURE (Learning from the Unseen: Robust Estimator). LURE is multiply robust, remaining consistent under several combinations of correctly specified nuisance components, and is asymptotically normal, enabling valid statistical inference. To our knowledge, this is the first work to address offline RL with hidden actions. We demonstrate LURE's effectiveness through simulations and a sepsis management application using the MIMIC-III database.

Referring multi-object tracking (RMOT) extends tracking from category-driven perception to language-guided understanding by grounding object trajectories in natural-language expressions. Despite recent progress, existing RMOT studies are largely conducted under in-domain settings, leaving the robustness of language-conditioned tracking under inevitable visual domain shifts unexplored. In this paper, we study Cross-Domain Referring Multi-Object Tracking (CD-RMOT), a new and challenging problem that evaluates whether an RMOT model trained on a labeled source domain can reliably follow natural-language expressions in an unlabeled target domain with different visual conditions. To support systematic study, we construct CD-RMOT-Bench, a unified benchmark that combines real clear-domain referring tracking data, aligned digital-twin variants, and real adverse-domain videos. CD-RMOT-Bench enables both controlled weather/viewpoint shift analysis and realistic synthetic-real transfer evaluation under a shared RMOT protocol. Further, we provide a Query-Centric Adaptation (QCA) framework, designed to stabilize the query space that bridges visual trajectories and referring expressions. Extensive experiments reveal that domain shifts severely degrade RMOT performance, where the failure is not merely caused by object detection errors but more critically by unstable expression-conditioned temporal association and target selection. QCA establishes a strong baseline, while CD-RMOT-Bench opens a new direction for robust language-guided tracking across visual domains.

Different research lines use the term world model in different ways, yet they share a common aim: to capture how the world evolves under action in a form that supports perception, simulation, and planning. Two prominent realizations are neural predictors that learn dynamics in continuous vector spaces, and hand-built physics engines that expose explicit state and physical laws. Neural predictors scale from data but leave the form of the dynamics implicit; physics engines are inspectable and editable but difficult to construct at scale. We introduce VisualPatchWorld (VPW), which represents world dynamics as code. VPW first selects a qualitative dynamical form with short active probes, then fits that form's free parameters from recorded state-action traces by minimizing multi-step prediction error. The resulting programs can be rolled forward like a simulator, inspected in source form, and used inside model-predictive control; image-derived scene graphs can supply the live state at replan time. Across comparisons with prior code-based world models, VPW attains 69.0% mean planning success and exceeds the strongest code baseline by 23.5 points. The largest gains arise when choosing the correct qualitative dynamics is essential. Under the same planner, the induced models approach ground-truth engine success on navigation and grasp-rich control; a residual gap remains for contact-rich pushing, and checking a shortlist of promising plans in the engine closes most of that gap. These results establish a practical route toward automatically constructed code world models that are useful for planning. Code is available at https://github.com/HKBU-KnowComp/VisualPatchWorld/.

Stereo depth estimation for driving, robotics and augmented reality must run at high resolution under tight latency budgets, yet in transformer-based matchers the global self-attention that aggregates scene context grows quadratically with the number of pixels and comes to dominate runtime. We show that the joint self-attention stage of a stereo transformer, whose role is to spread context across both views, can be replaced by a data-independent Walsh-Hadamard token mixer that mixes tokens globally in the transform domain at log-linear cost, while the data-dependent cross-attention that performs left-right correspondence is retained. On synthetic driving data the mixer matches the attention baseline in end-point error while reducing model compute by a factor of 2.46 and single-image inference latency by a factor of 2.65. A complexity analysis shows the benefit is governed by the ratio of sequence length to channel width, which explains why high-resolution stereo matching is a particularly favorable setting and why classification transformers are not; we confirm this token-to-channel scaling on non-stereo long-sequence benchmarks. Furthermore, we introduce a hybrid log-disparity loss function designed to up-weight small-disparity pixels corresponding to long-range objects. This approach reduces the error on distant objects without incurring any additional computational overhead.

Digital phenotyping (DP) using smartphones and wearable devices has shown considerable potential for mental health monitoring. However, progress remains difficult to evaluate due to heterogeneous datasets, inconsistent preprocessing pipelines. In this study, we present a reproducible benchmark built upon the Neurai-VN dataset, a high-resolution, multimodal dataset comprising passive sensing and active assessment from wearable and smartphone devices, collected from 100 Vietnamese adults over two weeks. The benchmark defines four clinically relevant binary classification tasks evaluated using standardized subject-wise cross-validation. Representative linear, tree-based, and neural baseline models are evaluated across predefined feature configurations. Mean subject-level F1 scores across five cross-validation folds reached 0.71 for Healthy Control vs. Depression and Healthy Control vs. Clinical, while Healthy Control vs. Anxiety and Depression vs. Anxiety achieved 0.69 and 0.56, respectively. These benchmark results provide reproducible baselines for future research on multimodal DP for mental health classification tasks.

Large Language Models (LLMs) have been widely applied in high-stakes decision-making scenarios such as corporate strategy, and users are increasingly relying on their outputs. However, the deep integration of open-source model sharing ecosystems with LLM-powered critical decision-making applications also introduces critical risks: if an attacker can manipulate the model's cognitive stance, they can indirectly influence the judgments and actions of downstream decision-makers. This paper defines such threats as decision-level hijacking. Existing attacks fail to achieve targeted cognitive manipulation without triggering prohibited content or degrading model functionality. To fill this gap, this paper reveals that Bit-Flip Attacks (BFAs) can serve as an attack vector for inducing decision-level hijacking, requiring no real-time interaction or control over the training process, and only a minimal number of weight bits need to be flipped after deployment to achieve stealthy, low-cost, and persistent cognitive manipulation. Therefore, we propose CogBias, a cognitive bias injection framework for LLMs. CogBias converts subjective preferences into optimization signals via a differentiable sentiment evaluator, uses a multi-objective loss to jointly constrain multiple dimensions, and constructs BitScout to locate critical bits, achieving targeted cognitive intervention under an ultra-sparse flip budget. Experiments on Llama-3.2-3B, Mistral-7B, and Qwen2.5-14B, as well as on the commercial recommendation and controversial factual topic scenarios, demonstrate that flipping only a small number of bits stably induces significant stance shifts on target topics, while the impact on non-target tasks and overall output distribution is limited. This work demonstrates that minute perturbations to low-level weight data suffice to undermine the high-level value alignment of LLMs.

LLMs are increasingly used for code generation in critical infrastructure, yet the security effect of domain-specific prompting is understudied. We present SecDrift, a benchmark measuring sector-conditioned security drift: the change in static-analysis vulnerability rates when prompts are conditioned on industry contexts versus neutral baselines. We evaluate 7 LLMs (6 producing analyzable code) across 8 CISA critical infrastructure sectors and 9 CWE categories with 5 replicates (5,355 evaluations), using a 5-dimension transformation with a matched-baseline condition that holds the task fixed while substituting only domain terminology. Industry prompts naively appear more secure (14.0% vs. 11.4%, -2.7pp), but the gap is not statistically significant (Fisher's exact p = 0.24, Cohen's h = -0.08) and is a composition artifact of two CWE categories: excluding CWE-502 and CWE-22 eliminates and slightly reverses it (+0.4pp, p = 1.00). A mixed-effects logistic regression confirms sector identity is not a moderator and localizes the only detectable condition effect to those two vulnerability types. 0 of 8 sectors show drift distinguishable from baseline, corrected or uncorrected (|h| < 0.15). A placebo on two non-CISA sectors (e-commerce, online education) reproduces the CISA industry rate almost exactly (10.5% vs. 11.4%, p = 0.63): the small pooled pattern reflects generic industry-framing specificity, not critical-infrastructure identity. In contrast, model selection has a large and consistent effect: among full-output models vulnerability rates range from 11.6% to 16.1%, and these differences persist across conditions. Model choice, not prompt framing, is the more reliable security lever. We release the framework, prompts, generated code, findings, human-validation verdicts, and analysis scripts.

Social navigation has progressed from simplified 2D environments toward a more general vision-based setting, in which a robot needs to achieve socially compliant behavior purely from onboard visual observations. Yet supporting simulation platforms have not kept pace: existing options either lack visual observations, lack moving human avatars, or fall short of real-world fidelity in appearance and pedestrian behavior, offering limited support for advancing vision-based social navigation. We introduce SONG, a SOcial Navigation platform powered by 3D Gaussian splatting (3DGS). It leverages 3DGS for both scene and avatar representations, drives pedestrians using semantically grounded trajectories generated by a large language model, and synthesizes their full-body motion with a trajectory-conditioned generator to produce continuous, natural movement. On top of the platform, we curate SONG-Bench, a set of evaluation episodes stratified by difficulty, and propose a multi-dimensional metric suite covering effectiveness, safety, and social compliance. A systematic evaluation of representative navigation baselines reveals three findings: (a) vision-based social navigation is far from solved; (b) a critical safety deficit precedes social etiquette; (c) real-world data matters more than model scale. Crucially, we demonstrate that fine-tuning on our curated data effectively improves the success rate in real-world environments. We hope our platform provides a faithful and rigorous testbed for the next generation of vision-based social navigation research.

Oblique-view urban remote sensing imagery inevitably exhibits geometric projection displacements between building roofs and footprints, leading to significant distortions in spatial structure. Existing approaches either ignore these deformations or handle them implicitly within segmentation-based frameworks, where progress is dominated by general segmentation advances rather than improvements in geometric correction. In this work, we explicitly define roof-to-footprint offset vector (RFOV) extraction as an independent learning task that decouples geometric alignment from semantic segmentation. To support this task, we introduce the Oblique City dataset (ObliCity), the first large-scale benchmark that integrates high-resolution UAV imagery and globally distributed satellite data, covering diverse city morphologies and camera perspectives. Methodologically, we reformulate DragOSM into DragRoof, an ODE-based framework inspired by human annotation behavior. By simulating the continuous process of dragging roofs toward their footprints, DragRoof learns deterministic, geometry-consistent offset fields and adaptively determines convergence through an end token. Extensive experiments on ObliCity demonstrate that DragRoof achieves state-of-the-art RFOV extraction performance, requiring fewer inference steps while delivering superior directional and length accuracy. Our dataset and model establish a principled foundation for studying projection displacement correction in oblique remote sensing imagery. The source code and dataset will be avaliable at https://github.com/likaiucas/DragRoof.

This paper investigates a hybrid reinforcement learning setting in tabular Markov Decision Processes (MDPs), where an agent aims to learn an optimal policy by combining online interactions with a target environment and offline data from a source environment. A central challenge is that offline data may be collected from outdated environments with shifted transition dynamics, making naive integration of historical data ineffective. To address this, we propose a unified algorithmic framework featuring two algorithms: MIN-UCB-VI for regret minimization and MAX-LCB-VI for best policy identification. Both algorithms leverage fine-grained bias information to more effectively exploit offline data under general transition shifts. We provide theoretical guarantees for our framework, including both instance-dependent and independent upper bounds on regret and sub-optimality gap. Furthermore, we establish matching lower bounds to demonstrate the optimality of our approach and validate our theoretical findings through extensive experiments.

Sleep staging remains challenging due to long-range temporal dependencies, ambiguous stage transitions-particularly in N1-and substantial distribution shifts across subjects, sampling rates, and EEG montages. These difficulties are further amplified in single-channel, low-latency scenarios required by wearable and real-world applications. To address these issues, we propose LGFNet, a CTC-guided sequence-to-sequence framework for robust sleep staging. LGFNet introduces a Local-Global Fusion encoder that jointly models fine-grained temporal dynamics and long-range sleep structure, overcoming the limitations of conventional serial hybrid architectures. A CTC-Attention joint training paradigm is adopted to unify temporal alignment with context-dependent modeling, enabling more accurate recognition of stage boundaries and transitions. Furthermore, a three-stage decoding strategy is devised, leveraging CTC-guided decoding and Viterbi-based smoothing to reduce error accumulation and enforce physiological consistency. Extensive cross-dataset evaluations on five public benchmarks demonstrate that LGFNet consistently outperforms state-of-the-art single-channel methods. In particular, on Sleep-EDF-78, LGFNet surpasses DMIN by +1.27% accuracy, +1.74% macro-F1, and +1.93% kappa, with pronounced gains on N1 and transition segments, highlighting its robustness and strong generalization across diverse sampling rates, montages, and recording environments.

Contrastive decoding (CD) has been proposed as a training-free strategy for mitigating object hallucinations in multimodal large language models (MLLMs), with reported gains on benchmarks such as POPE. However, recent work has questioned whether these gains reflect genuine improvements in visual grounding. In this study, we reproduce and extend the findings of "The Mirage of Performance Gains: Why Contrastive Decoding Fails to Mitigate Object Hallucinations in MLLMs." Specifically, we test the claim that CD induces a unidirectional output distribution shift in discriminative datasets and examine its generalizability across datasets. We also verify that the adaptive plausibility constraint (APC) reduces sampling to greedy search on both discriminative and generative benchmarks. Beyond reproduction, we rigorously study the effects of CD across generative and discriminative datasets. We conduct several experiments that provide additional insights: we analyze the logit distributions induced by different CD strategies on generative datasets, propose a proxy method and compare its performance against CD techniques, and investigate how hallucination signals propagate through each layer of the expert and amateur models. Experimental results across MME, POPE, and CHAIR using LLaVA and Qwen validate the original claims and show that the apparent improvements from CD are often spurious and do not consistently translate into stronger visual grounding for reducing hallucinations. These findings challenge the effectiveness of current contrastive decoding strategies and motivate the development of more reliable approaches for mitigating hallucinations in MLLMs.

Background: Most medical large language model (LLM) benchmarks focus on examination knowledge or isolated tasks and may not reflect the longitudinal, multimodal, and safety-critical workflow of cardiovascular care. Objective: To develop MyoCardBench, a real-world benchmark spanning the cardiovascular care continuum, and assess LLM performance across clinical dimensions and specialist tasks. Methods: MyoCardBench includes 2,263 items from 13 task-specific datasets derived from de-identified cardiovascular records and examination data. Sixteen cardiology physicians conducted annotation and reference construction, followed by cross-review from two senior cardiologists. Seven LLMs generated 15,841 outputs under standardized zero-shot settings. Open-ended tasks were evaluated using key-point coverage and holistic clinical quality, while CardioEthics was scored by accuracy. Results: GPT-5.4 achieved the highest macro-average (62.55) and item-weighted mean (62.19), followed by Gemini 3.1 Pro (59.95) and Qwen 3.6 27B (59.72). GPT-5.4 ranked first in all three dimensions. CardioAuxReport performed best (86.38), whereas CardioECGRead (17.25) and CardioEthics (17.34) were lowest. The largest gaps between holistic clinical quality and key-point coverage occurred in CardioComm (52.71), CardioEmergRescue (52.05), and CardioTreatPlan (48.80). Conclusions: To our knowledge, MyoCardBench is the largest real-world, multi-task benchmark for LLM evaluation across the cardiovascular care continuum and offers the broadest coverage of clinically authentic cardiology scenarios reported to date. It provides a rigorous framework for identifying model strengths, clinically important omissions, and priorities for future development.

The ability to retrieve relevant tables for answering questions is a key task for structured information retrieval. Multi-stage retrieval systems rely heavily on rerankers to refine candidate lists produced by efficient first-stage retrievers. As a result, neural rerankers and LLM-based reranking methods have become increasingly important due to their superior capacity for semantic understanding and reasoning compared to conventional sparse or dense retrieval models. Recently, Large Reasoning Models (LRMs) equipped with explicit chain-of-thought (CoT) reasoning have shown strong improvements in ranking quality in unstructured passage retrieval. In this work, we present TabRank, a framework for training reasoning rerankers for Tabular Retrieval. We first present a comprehensive dataset of 6728 reasoning traces for tabular reranking on the Natural Questions Tables dataset. We then explore two variants of training a compact reasoning model on these reasoning traces: explicit CoT distillation and conditioning the student reranker on the teacher's reasoning trace within the prompt. We stress-test TabRank on several out-of-distribution generalization settings on diverse domains and multi-table scenarios. Our approach significantly improves performance across a variety of table retrieval datasets, increasing Acc@10 by 30.5% on HybridQA, 15.2% on SQA, 52.9% on TabFact, and 13.1% on TATQA subsets of the Multi-Table QA Benchmark compared to the base model. Notably, TabRank generalizes effectively to multi-table reasoning. Our code, data and models are available at https://github.com/AdarshSingh7647/TabRanker

Message-passing graph neural networks are bounded by the 1-WL test and can miss topological structure that distinguishes non-isomorphic graphs. Positional and structural encodings (PSE) inject such topology-derived signals, and learned PSE encoders such as GPSE pretrain a single encoder to produce these signals from random node probes, which can then be frozen and reused as inputs across downstream graph models. We present CondPSE, a learned PSE encoder that applies a learnable polynomial graph filter bank to standard Gaussian node probes and refines the resulting structural-response branches through FiLM-style modulation conditioned on cross-filter, local message-passing, and graph-level signals. CondPSE is pretrained to reconstruct node-level positional/structural targets and graph-level invariants, and is then frozen for use as a downstream input encoding. On synthetic structural-discrimination benchmarks, CondPSE separates graph structures that 1-WL-bounded message passing cannot: it raises CSL accuracy from 42.9% to 97.3% and EXP accuracy from 68.3% to 99.9% relative to GPSE, and ablations show that the polynomial filter bank accounts for most of this gain. On real molecular property prediction, the picture is more limited. With a hybrid local-message-passing/global-attention backbone, CondPSE performs comparably to GPSE without surpassing it, and a ZINC backbone sweep shows no consistent ordering between the two encoders. We report these results and discuss why strong synthetic structural discrimination does not, on its own, yield a downstream advantage for frozen learned PSE encoders, including the role of downstream integration and possible mismatch between structural pretraining targets and molecular property labels.

Language models have demonstrated remarkable success in solving a wide range of tasks. However, answering complex scientific questions about the power flow often requires solving the distribution optimal power flow (D-OPF) problem. These questions call for numerical solvers and simulators, as linguistic reasoning from parametric knowledge often gives incorrect answers. In this work, we present VeraGrid-Agent, a tool-augmented LLM that autonomously writes the simulator input, executes the open-source VeraGrid solver, and reads the solver output before answering. To evaluate performance, we introduce VeraGrid-MCQ-150, a set of deterministic, expert template driven, $150$ multiple-choice questions. We evaluate the performance under two regimes: (i) no-tool reasoning and (ii) agent (LLM with simulator access). Without tools, every model performs with an accuracy of $42.7\%$--$49.3\%$. However, with VeraGrid-Agent, accuracy increases to $97.3\%$--$100.0\%$. We also do a failure-mode analysis to show that the few remaining errors arise from wrong interpretations during multi-step reasoning, rather than any failure in the simulators execution.

The distance between persons reveals significant information about their perception of each other. However, such information is not easily extractable and interpretable from video input. We developed an open-sourced library, Facial Interpersonal Distance Analysis and Coding (FIDAC) that transforms facial detection results into actionable data about location and interpersonal distance. This tool merges data from multiple open-source facial detection models, strategically compensating for gaps in any individual model. In addition, we include methods for more accurate tracking, such as a pipeline for human coding of the selection of faces and a benchmarking tool to reduce depth distortion. For next steps, we plan on building upon FIDAC by evaluating its effectiveness at measuring interpersonal distance at various depths and orientations while further integrating features of proxemic analysis such as synchrony into its software.

This paper studies the behavior of language models in a multi-agent crowd simulation, focusing on how affect propagates among agents that perceive and appraise one another. Each agent perceives its neighbors through visual, auditory, and tactile channels, then appraises these perceptions in light of its prompted personality profile, memory, current affective state, and situational context. Appraisal is carried out by an LLM, which updates the agent's internal affective state and selects its outward expression. The architecture contains no hand-authored mechanism for directly transferring affective state between agents; instead, inter-agent influence arises through the perception-appraisal-expression loop. The agent representation draws on the Big Five personality model and Russell's circumplex model of affect. To limit latency, low-level steering and navigation are handled by a conventional crowd simulator operating independently of the LLM-based cognitive layer. We evaluate the architecture across five scenario environments spanning alarming, joyful, and neutral situations in different spatial layouts. The results show that the system produces emotional contagion dynamics with spatial, temporal, and personality-dependent structure in sparse, small crowds. Alarm spreads from seeded agents as a traveling front, the mean alarmed fraction settles at a nonzero plateau, and the distribution of prompted personality profiles determines whether an ambiguous alarm ignites panic and whether a provocation is interpreted as anger or fear. We further evaluate the appraisal step through controlled experiments across prompt variants, sampling temperatures, and four model backends, showing that the dynamics are backend-dependent.

Safety inflation can cause nearby obstacles to overlap, violating the disjoint-obstacle assumptions used by many modulation-based reactive planners. We investigate Star-World workspace reshaping for three-dimensional reactive control of a Franka Emika Panda manipulator. At each update, intersecting inflated obstacles are clustered and replaced by star-shaped proxies before a dynamical-system-based end-effector controller is evaluated. A null-space artificial-potential-field term provides complementary arm-body avoidance. We compare reshaped and unreshaped obstacle representations in six PyBullet scenarios using goal attainment, path-length ratio, and computation time. In this preliminary 12-trial evaluation, reshaping reaches the goal in five of six scenarios, compared with four of six for the unreshaped baseline. It resolves the canonical overlapping-wall case and requires 0.68--8.70\,ms per workspace update for scenes containing one to seven obstacles. However, it also increases path length, produces near-equilibria in two cases, and closes a navigable corridor through over-aggressive merging. These results show both the promise and the practical limitations of transferring Star-World guarantees from workspace geometry to a redundant manipulator controlled through inverse kinematics.

Recent advances in RAG aim to optimize for performance by paying high ingestion costs for knowledge ingestion: building knowledge graphs or extracting SQL tables. In this work we show that the operations that such knowledge bases allow can be replicated with zero ingestion costs (not even a vector database); in fact our solution, Zero-Ingestion ScalableRAG, handily out-performs all baselines (including knowledge graph approaches) in three out of the six corpora considered here, and only marginally missing maximum performance on the other three, with average accuracy across all six datasets 7.36% above the next most competitive baseline. It achieves this by keeping a workspace of document sets and values sets that it can write into and read from, allowing for on-the-fly aggregative reasoning in all situations where grouping is required on a primary key that is in one to one correspondence with a subset of the total document set. Capping the number of LLM calls by a constant independent of the corpus size, we also introduce Limited-Ingestion ScalableRAG, which does use a minimal vector database as well as an automated pattern discovery from a sample of documents, to further improve accuracy at scale. Our code is available at https://github.com/cohesity/ScalableRAG .

A central challenge in interpreting learned decision-making systems is to determine whether their internal representations contain concepts that help explain their behavior. We report interpretability experiments for a tokenized autoregressive Transformer agent in the Game of Hidden Rules (GOHR). We focus on a compact two-rule task in which both hidden rules map object shapes to target buckets, but with different permutations. The policy is trained on episodes sampled from these two hidden rules and then evaluated with fixed weights. It is never given a rule label and does not use an explicit rule classifier; any rule information must be inferred implicitly from interaction history. In this setting, the correct rule is not identifiable before the agent tries an informative move and observes accept/reject feedback. Sparse autoencoders (SAEs) trained on the agent's decision-token embeddings recover this structure. When held-out decisions are labeled by simple concepts such as the chosen shape or bucket, SAE dimensions that are highly selective for a concept cover most decisions where that concept is present. Individual SAE dimensions also correspond to interpretable strategies such as probing one rule hypothesis and switching after negative feedback.

Imperfections in AI-generated code require that software developers modify the generated code manually, or by re-prompting an AI programming assistant. Manual code edits provide more realistic and granular information on editing behavior than Git commits, which only contain final successful code snippets. Yet, due to a lack of high-quality, realistic code editing data, LLMs are mostly trained on publicly available Git data (e.g., commits). To address this gap, we introduce DECODE (Developer Edits of Code Dataset), a dataset of 53.6K real-world in-IDE code edits of AI-generated code in Python, TypeScript, and JavaScript, sourced from 1K+ developers. First, we demonstrate the utility of DECODE for data analysis, obtaining insights on when, why, and how AI-generated code is edited. We find that most edits occur within the first 15 minutes after accepting an AI completion, resulting in the removal of AI completions in 31% of edit trajectories. Second, we use DECODE to benchmark the ability of LLMs to predict code edits. We find that finetuning on DECODE enables open-source 3B models to perform code edit prediction tasks significantly better than frontier LLMs. We then discuss implications of this work, emphasizing the necessity of developer-centric machine learning approaches for future AI programming assistants.

Despite rapid progress in Multi-modal Large Language Models (MLLMs), understanding long-form videos is still bottlenecked by limited context windows. While recent keyframe sampling methods attempt to mitigate this by distilling video inputs into a compact set of query-relevant frames, navigating the vast spatio-temporal search space remains challenging, as spatial detail and temporal coverage often conflict. To address this, we introduce LENS, a training-free keyframe sampling framework that dynamically decides when to zoom in for fine-grained details and when to zoom out for broader context based on the text query. Concretely, LENS adaptively allocates a limited frame budget between spatial zoom-ins, which highlight query-relevant regions within individual frames, and temporal zoom-outs, which expand the temporal scope through multi-frame aggregation, enabling the model to reason across multiple granularities while capturing both high-fidelity details and long-range context. Across diverse long-form video benchmarks, LENS consistently outperforms prior state-of-the-art keyframe sampling methods and delivers substantial gains over uniform sampling, improving Video-MME accuracy from 53.3% to 60.7% with Qwen2.5-VL.Code is available at https://github.com/zhangce01/LENS.

Experience replay remains one of the most practical and useful algorithmic tools in the deep reinforcement learning (DRL) toolbox. Aside from the limited success of prioritized replay and specialized approaches for large asynchronous systems, most DRL algorithms make use of a large, uniformly sampled recency buffer---even the size, one million, remains unchanged. Could we store less data, reduce redundancy, or more effectively chain experience together to speed up value propagation and still retain the performance of large buffers? In this paper, we investigate a simple compression approach that stores representative transitions derived from the end-points of a chain of connected $n$-step sequences. By curating these end-points in a smaller recency buffer, our method maintains an effective memory horizon comparable to a standard large buffer while requiring an order of magnitude less storage. Through empirical evaluation, we demonstrate that this approach prevents the systematic bias inherent in naive compression strategies and matches the performance of traditional large buffers in the Pinball environment and the Atari 2600 benchmark.

Early ranking stages in recommendation systems precompute item embeddings and cache them in-model for scoring within strict latency constraints. Because this cache exists only at serving time, outside the training loop, training and serving use different item representations, a structural discrepancy that limits quality and adds operational fragility. We show that co-designing the training and serving paths removes this representation discrepancy at its source. We introduce the memory layer, an in-model key-value embedding cache co-trained with the model: the item tower writes embeddings during training and the model reads them at serving, one source of truth for item representations by construction. Always-on embeddings cover items not yet cached, so every item receives a prediction, and the design consolidates three separate trainer-to-predictor update paths into a single self-contained pipeline. Deployed in production on Instagram Reels, the memory layer raises prediction coverage from 96% to 100%, improves embedding freshness from $O(5\text{ min})$ to $O(20\text{ s})$, and narrows the training-serving Normalized Entropy (NE) gap by up to 86%, yielding over $2\times$ recall for the freshest content and a 5-6% cold start engagement lift. Because embeddings are produced during training, the system needs no separate bulk-evaluation or publish-time recomputation, cutting training-and-publish computational cost by 30% at neutral serving computational cost.

Biomedical image analysis spans diverse modalities and tasks, yet real-world deployment is hindered by severe distribution shifts across scanners, protocols, and patient populations. High-performing models consequently require repeated domain-specific fine-tuning, which is a costly cycle that becomes impractical when labels are scarce or privacy constraints limit data sharing. We propose OPERA (Offline Policy-guided Expert Routing and Adaptation), a multi-agent ensemble framework that addresses this deployment bottleneck by treating expert weight assignment as an offline policy learning problem: a routing policy is learned from a small validation set without gradient updates to any expert agent, then deployed with test-time adaptation to handle distribution shift. OPERA coordinates heterogeneous specialist agents through complementary mechanisms. The expert profiling module learns selection policies offline, enabling informed allocation of expertise. Each agent undergoes confidence calibration through temperature adjustment, ensuring more reliable probabilistic outputs. OPERA also incorporates distribution aware adaptation, where class weights are dynamically adjusted at the batch level using statistics derived from unlabeled test data. Instance level routing assigns each sample to the most suitable expert by leveraging inter model agreement and predictive entropy. We evaluate OPERA on 9 datasets covering fundus photography, chest X-ray, CT, MRI, and multimodal diagnostic benchmarks, comparing against 30+ baselines across classification, segmentation, and multimodal settings. OPERA consistently improves performance and calibration quality, demonstrating that offline policy-guided expert agents coordination is a practical path to deployable biomedical AI without retraining. Code is on \href{https://github.com/HUANGLIZI/OPERA}{GitHub}.

Embodied AI increasingly relies on queryable semantic maps built from pre-trained vision-language models to enable zero-shot Object Goal Navigation (ObjectNav). However, existing approaches typically depend on text-only queries, which become less reliable as semantic specificity increases toward fine-grained object categories. We introduce IMPRINT, a zero-shot plug-and-play framework that enriches textual object queries with web-sourced images to improve grounding in queryable maps. Retrieved images are encoded using a vision-language model, matched against the semantic map to produce similarity maps, and aggregated to yield context-aware localization. Notably, this requires no training or modification of the underlying navigation policy. To explicitly evaluate long-tail behavior, we present HSSD-rare, a new ObjectNav benchmark built on Habitat Synthetic Scenes and featuring semantically specific subcategories. Across both OVON and HSSD-rare, image-conditioned queries consistently improve object grounding and yield end-to-end navigation gains. Further analysis reveals that translating localization gains to navigation performance depends critically on downstream detection quality, highlighting a key systems bottleneck in long-tail embodied navigation.

Long-horizon LLM agents accumulate reasoning traces, actions, and tool observations that can eventually exceed a model's fixed context window. Existing compaction methods address this limitation by discarding, summarizing, or retrieving earlier information, but they may remove task-critical details or fail to recover them reliably. We propose ARC (Addressable Recall Compaction), a context-management framework that separates archival storage from active-context presentation. ARC stores tool observations in an append-only, ID-addressable log and replaces older observations with compact citations when compaction is required. The agent can subsequently use these identifiers to request stored content without re-executing the corresponding tools or depending solely on similarity-based retrieval. We evaluate ARC using Qwen3-8B with a 16k context window and Qwen3-32B with a 32k context window. On the Needle-in-a-Haystack evaluation, ARC achieves an average exact-answer accuracy of 99.40%, compared with 88.12% for the best-performing baseline in our evaluation. ARC also reduces estimated serving time and HBM traffic under our hardware-cost model. On the LongBench-v2 Hard subset, ARC obtains an average accuracy of 29.97%, compared with 28.25% for the best-performing baseline. These results indicate that explicit, address-based recall can improve information retention and serving efficiency relative to the evaluated context-management baselines under the tested settings.

We show with experiments and system-level simulations that it is possible to successfully mitigate the impact of retention loss on inference accuracy degradation by using both circuit-level compensation techniques and batch normalization recalibration at the algorithmic level. Experiments are performed on a single-poly floating-gate (FG) analog non-volatile memory array for analog in-memory computing fabricated in a standard 65 nm CMOS. We use a model of retention-loss statistics calibrated with experiments to evaluate the system-level impact on neural network models such as VGG-10/CIFAR-10 and WideResNet-28-10/CIFAR-100. We show that, after 60 days since programming, combined mitigation techniques enable to recover the baseline inference accuracy within 2-4%

Classical grasp quality metrics assume a single deterministic friction coefficient, so they cannot predict whether a grasp retains force closure across the range of friction values the contacting surfaces may exhibit. To predict these failures, we present FIRMGrasp, a family of friction-volatility-aware grasp quality metrics grounded in the Conditional Value-at-Risk (CVaR) risk measure. Unlike standard grasp quality assessors that assume a single friction realization, our metric evaluates the force-closure margin at the CVaR-discounted mean of the adverse friction tail, yielding a risk-adjusted margin $\varepsilon^{(β)}$, the inscribed-ball radius of the risk-adjusted wrench space. We establish its monotonicity in the confidence level $β$, its differentiability in the grasp parameters, and a probabilistic closure certificate that guarantees force closure with probability at least $β$ whenever $\varepsilon^{(β)}$ is positive. Under a calibrated friction distribution, analytic evaluation shows our $\varepsilon^{(β)}$ metric identifies friction-sensitive grasps that the nominal Ferrari-Canny epsilon rates as high-quality, and we compare against the nominal epsilon and recent differentiable baselines. Across 1,599 LEAP Hand and Allegro Hand grasps, 53% of the grasps the nominal Ferrari-Canny margin certifies lose force closure in the adverse friction tail. On the same set, the nominal margin separates realized shake and pick success with probabilities of only 0.53 and 0.67, near chance on shake success, whereas $\varepsilon^{(β)}$ orders the pair correctly with probabilities of 0.63 and 0.78, respectively. In simulated lift trials with gravity enabled at an adverse friction coefficient of 0.2, grasps $\varepsilon^{(β)}$ certifies reach a 70% success rate under lateral pull, against 25% for grasps the nominal margin certifies but $\varepsilon^{(β)}$ rejects.

A growing number of applications, such as biometrics and retrieval-augmented generation (RAG), rely on cosine similarity scores computed between vector embeddings of text, images, or audio. These systems return similarity scores through their APIs for ranking and verification. However, such releases can leak information about individual records and enable membership inference attacks. While differential privacy (DP) provides a principled metric for quantifying attack risks, naïve application of DP mechanisms---such as adding i.i.d. Gaussian noise to vector entries---leads to excessive distortion (i.e., low utility) at a given privacy constraint that scales poorly with the number of released scores. We propose \textsc{ScoreShield}, a perturb-then-project mechanism that adds Gaussian noise calibrated to global sensitivity of the chosen score release regime and then projects the result onto the feasibility set of valid cosine objects. \textsc{ScoreShield} satisfies $(\varepsilon,δ)$-DP for releasing similarity score vectors and Gram matrices. We provide utility guarantees for the exact Frobenius metric projection used in the risk analysis, and prove convergence to feasibility for the practical averaged alternating-projection solver used for large-scale Gram releases. For full pairwise cosine Gram release under record-level replacement adjacency, the exact-projection bound improves the $n$-dependence of squared Frobenius risk from $Θ(n^3)$ for the naïve Gaussian baseline to $\mathcal{O}(n^2)$ for fixed privacy parameters, with sharper local bounds at low-rank Grams. We evaluate the mechanism across RAG, face recognition, semantic retrieval, image similarity, and recommender-system tasks.

LLM cognitive scores are increasingly summarized as per-ability profiles whose dimensions should converge across tasks, respond selectively to matched interventions, and generalize beyond the models used to define them. We introduce CogArena, a procedurally generated 13-paradigm benchmark built around a multimethod framework for determining when cognitive-task scores warrant dimensional labels across five theory-motivated groupings. Across 55 open-weight models, nearly all paradigm correlations are positive and a common axis explains about half the variance. The within-grouping advantage is small, scoring-sensitive, and uncertain across model families. In a separately frozen, fully crossed study across 12 models from six families, targeted scaffolds show a small matched-grouping advantage, but no scaffold-specific contrast survives multiplicity correction and selectivity does not improve held-out-family prediction. The frozen confirmation criterion fails. A post-hoc alternate-wording replication produces a smaller positive estimate and again fails. Together, these results support a boundary conclusion. Theory-aligned prompting produces a small in-battery diagonal tendency, but the present evidence does not establish stable five-dimensional profiles. CogArena provides a workflow joining behavioral signatures, covariance, matched interventions, and out-of-family prediction before cognitive labels are attached to model scores.

Neural networks are hindered by accumulating dormant neurons and loss of expressivity throughout training, particularly in non-stationary data settings, such as continual supervised and reinforcement learning. Recently, neuron resets have been used to maintain gradient flow and restore plasticity. However, full unit reinitialization often sacrifices peak performance and can destabilize training, leading to policy collapse. To preserve plasticity without destabilizing training, we propose Calibrated Partial Resets (CPR), an optimizer that periodically pulls low-utility neurons toward their initialization, with pull strength scaled by each neuron's utility. Unlike binary reset methods, partial resets avoid brittleness; unlike uniform decay, calibrated utility-scaling concentrates adjustment on the units that need it most. Among compared methods, only CPR avoids policy collapse over 400M training steps in SlipperyAnt, and it outperforms prior decay and reset-based methods on Continual MetaWorld and Continual MinAtar benchmarks. Ablations reveal a tunable trade-off between plasticity and peak performance, highlighting utility-scaled reinitialization as a promising direction for continual learning.

Differentiable simulation can accelerate contact-rich trajectory optimisation by exposing local sensitivities of task outcomes to controls. Existing approaches either use finite differences, which are expensive and step-size sensitive; differentiate iterative contact solvers by unrolling automatic differentiation (AD), which stores a growing computation trace; or require intricate, solver-specific KKT sensitivity derivations. We introduce an AD-assisted implicit derivative for regularised smooth contacts and apply it to Mujoco MJX, based on the Implicit Function Theorem (IFT). The method differentiates the stationarity residual at the tolerance-converged solution, avoiding both solver unrolling and hand-assembled KKT systems. IFT keeps compiled temporary memory nearly constant with solver effort, changing by less than 4$\%$ from one to ten iterations versus 10.6$\times$ growth for unrolled AD. IFT memory grows slower with active contacts and model dimension, using 20$\times$ less memory at 256 contacts and 6$\times$ less at 16 contacts and 96 DoF. We further introduce optimiser distillation for residual MPC, amortising batched full-horizon iLQR into a policy that guides short-horizon residual iLQR. Across Finger, Franka, and Unitree, this raises six-step success by 28-98 percentage points over standard iLQR.

We introduce PerceptionBench, a benchmark specifically designed to evaluate the atomic visual perception capabilities of Multimodal Large Language Models (MLLMs). Existing benchmarks often fail to isolate perception: holistic evaluations conflate perceptual errors with failures in reasoning or domain knowledge, while application-driven benchmarks only cover narrow, fragmented domains shaped by heuristic designs. To address these limitations, PerceptionBench adopts a bottom-up approach: by diagnosing the earliest failure points in the responses of frontier MLLMs across 42 existing benchmarks, we construct an error taxonomy whose perception branch defines ten atomic perceptual capabilities. Guided by this taxonomy, we construct 3,000 verified questions with short, unambiguous answers, each isolating a single capability, with difficulty stemming from perception rather than reasoning or knowledge. Benchmark results across sixteen frontier MLLMs reveal that atomic perception remains largely unsolved---no model reaches 60\% accuracy, perception-related hallucination is the weakest capability on average, and similar overall scores conceal sharply divergent capability profiles. PerceptionBench thus provides a capability-level standard for measuring and diagnosing the visual perception boundaries of MLLMs.

This work proposes an adaptation of the attention mechanism for triangle meshes. The core observation is that endowing the attention mechanism with critical properties for learning over meshes -- intrinsicality and triangulation-agnosticism -- enables it to attain state-of-the-art results over several learning-based tasks in geometry-processing. The above is achieved by modifying the attention mechanism from the bottom up based on simple principles from geometry-processing. Namely, the quantities used within attention -- queries, keys and values -- are created by an intrinsic, triangulation-agnostic network, and treated as discretizations of continuous functions. From that, we devise an appropriate attention mechanism that operates over triangle meshes through standard FEM discretization of the resulting integrals of the above functions. Surprisingly, as far as we know, this straightforward approach has not been utilized for learning over meshes. Experiments show our method exceeds current state of the art, including both mesh-based architectures as well as point cloud transformers. Namely, we show significant improvements on several common benchmarks and tasks -- predicting canonical high-frequency signals; predicting deformations; computing dense correspondences, both between full shapes and partial ones; and predicting feature descriptors.

Multimodal foundation models learned to see and to speak by consuming the whole internet. Embodied agents admit no such shortcut, since they require data that couple observations with physical states and actions. These signals can be provided, to varying degrees, by multiple data sources. In this work, we organize the embodied data ecosystem as a "pyramid" spanning five complementary sources: real-robot data, UMI-style data, egocentric and exocentric data, simulation data, and general vision-language data. We organize the pyramid around the tension between scalability and robot alignment, and further characterize each source in terms of data quality, diversity, reusability, and physical fidelity. We then analyze recent embodied foundation models through the lens of their data recipes, examining how different sources are selected, aligned, and mixed during pretraining. For embodied brain models, vision-language-action models, and world-action models alike, we relate data composition to capabilities in perception, reasoning, planning, action generation, and world prediction. We close by discussing six open challenges: building large-scale tactile datasets, collecting failure and recovery data, developing scalable data-collection pipelines, aligning actions across embodiments, leveraging egocentric data for dexterous manipulation, and designing principled data recipes for robot learning. We hope this work paves the foundation for the design of next-generation embodied systems.

Multimodal large language models (MLLMs) hold immense potential to revolutionize clinical practice, yet deploying them in the medical domain is fundamentally a vision-centric challenge: models must absorb knowledge from heterogeneous 2D and 3D medical images, and evaluation protocols must align with radiologists' clinical practice and provide an accurate, fine-grained and factualness-driven assessment. In this paper, we introduce ClinFusion, a vision-centric MLLM designed for holistic medical understanding that systematically addresses these limitations. We propose a compositional and cascaded vision encoder architecture featuring a Cascade Spatial-Aware Locality Fusion operator that unifies diverse 2D and native 3D medical image understanding within a fused encoder. We further introduce a vision-grounded evaluation framework, including MedIF-Bench for instruction-following assessment and a region-of-interest-grounded method for clinically aligned and factualness-driven report generation evaluation. We show that ClinFusion sets a new state-of-the-art across a comprehensive suite of 2D and 3D multimodal medical benchmarks---spanning visual question answering, report generation, and instruction following---as well as textual medical tasks, outperforming leading open-source medical MLLMs (\textit{e.g.}, Hulu-Med, Lingshu) on 20 out of 24 benchmarks and demonstrating multimodal capabilities better than powerful proprietary models such as GPT-5.2 and Gemini-3-Flash on 13 out of 16 benchmarks, and can be further augmented with agentic tool use for retrieval-augmented and tool-assisted clinical workflows. A blinded evaluation by board-certified radiologists confirms that ClinFusion produces the highest-ranked reports, and validates our RoI-grounded metric as achieving the strongest correlation with expert judgment among all automatic evaluation metrics examined.

Computer vision models have become highly effective for medical applications, yet their black-box nature continues to undermine clinician trust. In clinical workflows, chest X-ray classifiers are increasingly paired with Vision-Language Models (VLMs) to generate natural-language explanations. However, these systems add linguistic fluency without addressing the underlying opacity of the visual model. With the emergence of Kolmogorov-Arnold Networks (KANs), whose spline-based components provide inherently interpretable functional units, we investigate whether this architectural transparency can be leveraged to produce more trustworthy textual explanations. We introduce KANEx, the first ever framework that leverages the symbolic transparency of KANs to ground VLM reasoning. This interpretability also made it possible to design KAN-Map, a novel heatmap generation method derived directly from KAN models rather than gradient approximations. We feed these grounded contexts into downstream VLMs for enhanced explainability. Benchmarked on the MIMIC-CXR dataset, we demonstrate that KAN-based architectures with ResNet/ViT baselines demonstrate improved semantic similarity while producing significantly more faithful saliency maps. KAN architectures improve visual localization and downstream reasoning quality by 10%. Our findings suggest that grounding linguistic explanations and visual attributions in mathematically interpretable units is a necessary step toward trustworthy medical AI.

Multi-turn long-horizon planning is critical for foundation model agents, yet how to fundamentally improve it remains unclear. Existing models are trained on uncontrollable and opaque Internet data, making it difficult to identify how planning ability is acquired, shaped, and integrated. To address this challenge, we introduce a unified and controlled multi-turn environment that enables precise control. It allows systematically study long-horizon planning across three stages. (1) Planning ability acquisition during pre-training. We study data format, distribution, and quality. Explicit world model construction through CoT state transition modeling yields stronger long-horizon generalization. Atomic skills alone are insufficient for compositional generalization, whereas a litte long-horizon data works. Moreover, suboptimal trajectories severely impair performance because errors amplify over long horizons. (2) Planning ability shaping via GRPO and OPD post-training. Through mutual information, we distinguish general planning patterns from task-specific planning knowledge. For planning patterns, we identify three application regions of post-training: unnecessary, effective, and unsupported. OPD has a broader effective region than GRPO under low-quality and long-horizon settings, as it provides more consistent update directions. For planning knowledge, distilling unseen procedures from a teacher with different knowledge may impair student's prior world modeling without fully establishing new knowledge. (3) Planning ability integration through MOPD post-training. We show that multi-teacher on-policy distillation (MOPD) integrates capabilities by converging to shared planning-pattern across environments. Compatible patterns enable cross-environment generalization, partially shared patterns support continual learning, while completely conflicting patterns cause severe interference.

Pretraining data processing is critical to the downstream performance of Large Language Models (LLMs). However, many existing approaches define a fixed processing strategy at the corpus or domain level and apply it uniformly to many examples, without adapting to the needs of each example. We propose DataOrchestra, a framework that unifies different processing operations and orchestrates an example-specific pipeline for each example. Given a chunk of pretraining data, an orchestrator decides whether to drop, untouch, or clean it. For a chunk to be cleaned, it selects one or more downstream operations, ranging from programmatic editing to different forms of LLM-based rewriting. For each rewriting step, it further generates a concrete instruction, which is executed by the corresponding downstream tool model. We pretrain models from 0.5B to 7B from scratch on web data processed by DataOrchestra and observe stable average gains over individual data-processing methods across 11 benchmarks. DataOrchestra is also effective for math continued pretraining and outperforms stronger processing baselines, while reducing processing compute by skipping unnecessary downstream operations.

Entity matching identifies records that refer to the same real-world entity. Language models can be adapted to this task through bi-encoder, cross-encoder, and generative matcher architectures. However, prior studies often conflate matcher architecture with differences in model backbone, model variant(reflecting different pretraining objectives), and model size, making it difficult to isolate the sources of performance gains. We address this issue through a controlled factorial study spanning three matcher architectures, three model variants and three model sizes from the Qwen3 family, and nine datasets, totaling 1,215 fine-tuning runs. We also evaluate cross-dataset transferability and computational cost. Our results show that model variant is critical for bi-encoders: embedding-oriented variants provide stronger initialization and more favorable representation geometry predictive of downstream matching performance. Cross-encoders retain a consistent advantage over bi-encoders because they jointly encode record pairs rather than representing each record independently, although larger models partially narrow this gap. Generative matchers do not universally outperform cross-encoders. Instead, their advantages concentrate under distribution shift, including subtle unseen differences in record schemas and cross-dataset transfer. We further find that larger models rely more heavily on shortcut learning and therefore do not necessarily perform better. These findings clarify the factors underlying performance differences across matcher architectures and motivate future research and benchmark designs that better disentangle architectural choices from model-level factors while explicitly evaluating distribution shift and cross-dataset transferability. We release our experimental results, code, training scripts, and evaluation data at https://github.com/Jantory/llm-trained-matcher.

We introduce Kimi K3, a 2.8T parameter Mixture-of-Experts model with 104 billion activated parameters, native vision capabilities, and a 1-million-token context window. Kimi K3 is built on Kimi Delta Attention and Attention Residuals, which improve information flow across sequence length and model depth. Together with Stable LatentMoE, which effectively activates 16 of 896 routed experts per token, and refined training and data recipes, these advances yield an approximately 2.5x improvement in overall scaling efficiency over Kimi K2. Post-training highlights reinforcement learning across general, agentic, and coding domains and multiple reasoning-effort levels, enabling compositional generalization and robust long-horizon execution. At 2.8T scale, Kimi K3 is supported by infrastructure advances in multiple areas: algorithm-system co-design for KDA, perfectly balanced expert-parallel training with efficient memory management, million-token agentic RL with persistent rollout and sandbox states, and deployment innovations. Extensive evaluations show that Kimi K3 achieves frontier-level performance across long-horizon coding, agentic, knowledge, reasoning, and vision tasks. While its overall performance still trails the most powerful proprietary models, namely Claude Fable 5 and GPT-5.6 Sol, Kimi K3 consistently outperforms other open and proprietary models evaluated in our suite. We release the full Kimi K3 model weights to facilitate future research and accelerate the broader deployment and adoption of frontier intelligence.

Reliable visual document understanding requires a model to attribute each answer to the evidence regions that support it. Recent benchmarks and systems express this step through a coordinate interface: the model outputs the coordinates of bounding boxes that mark the evidence regions in the document. Under this interface, vision-language models often fail to identify the right regions even when the answer is correct, a failure known as Attribution Hallucination. We present a study that investigates whether this failure is partially limited by what the model can express through coordinates. On a verified bilingual CiteVQA subset, we compare the coordinate interface with a language interface in which the model outputs only text, quoting its evidence verbatim, and a multimodal retriever returns the location of each quote as a page region proposed by a layout parser (tables and figures are quoted through their captions or notes); the comparison is repeated over six open vision-language models. Compared with the coordinate interface, evidence recall rises from at most 8 points to between 26 and 47 and the hallucination rate roughly halves, with little change in answer quality. Building on this comparison, we use the same quote-and-retrieve pipeline as a training scaffold: because region-level evidence labels are expensive to collect for long documents, we introduce a GRPO recipe whose reward is a judge's reading of the gold answer and crops of the retrieved regions, training the model to quote better evidence without any region labels and raising an 8B backbone's strict attributed accuracy from 22.4 to 33.8. These findings indicate a practical path to improve attribution"without a coordinate interface and without costly region-level supervision.

Generate--test--revise loops are common in coding agents, but repetition alone provides no reliability guarantee. We study the gap between finding a correct patch and retaining, verifying, and submitting it. A sealed five-seed study over 30 HumanEval repairs produces 900 three-revision trajectories. Under forced revision, current correctness with current traces falls from 0.820 after one revision to 0.673 after two, although ever-correct rises to 0.847. Two common-state studies use 2,430 branches from identical frozen programs to remove post-treatment risk-set bias. In a prespecified 14B replication, stale traces harm 34/135 correct starts versus 4/135 with current traces, a 22.2-point increase (task-cluster 95\% CI $[8.9,37.0]$, exact Holm $p=0.0337$). A prospective 540-rollout policy eliminates observed correct-start harm but reduces wrong-start repair and fails its joint criterion. Repository experiments over 24 bugs and four coder stacks expose floor effects and component heterogeneity without Holm-significant effects. We therefore separate admission, preservation, grounded certification, competence, and liveness. We derive an evidence-bound typed loop contract and instantiate its mechanically enforceable subset in a reference implementation that binds verifier evidence to exact code states, preserves verified checkpoints, and emits auditable admission receipts. The implementation is an executable specification and conformance artifact, not evidence of improved repair competence or calibrated verifier dependence.

Token-level sparse attention, as implemented by DeepSeek Sparse Attention (DSA) in production systems, makes the downstream attention efficient but shifts the bottleneck to the indexer that feeds it. To select the top-k tokens for each query, the indexer must still score every preceding token, incurring a cost of O(L^2) per layer for a sequence of length L. We observe that this per-query scan is largely redundant: nearby queries select highly overlapping top-k tokens, and the indexer scores are long-tailed along the key axis. We exploit these properties in PIVOT, Proxy Indexing Via One full-prefix Traversal, a training-free, drop-in replacement for the DSA indexer that shares one prefix scan across a group of nearby queries. PIVOT aggregates a group into a single proxy query, performs one shared full-prefix scan to obtain a candidate set, and then selects a top-k for each query from that set. Two variants trade speed for fidelity: PIVOT-Reuse shares the proxy top-k across the group for maximum speed, whereas PIVOT-Refine re-scores the candidate set with the indexer of each query and then selects an individual top-k, matching the dense indexer at a small additional cost. A single algorithm covers both inference phases, differing only in how groups are formed: fixed-size groups of consecutive queries in prefill, and the queries decoded together in one multi-token prediction (MTP) step in decode. On DeepSeek-V3.2 and GLM-5.1 across LongBench and RULER, PIVOT matches the accuracy of the dense DSA indexer while accelerating it by up to 4x and reducing end-to-end latency by up to 1.6x at long context.

Automatic Text Recognition (ATR) now supplies digital humanities with large volumes of unstructured, heterogeneous, and often noisy text in ancient languages. Downstream semantic analysestext reuse identification, alignment, and semantic search-rely on sentence embeddings, yet existing methods transfer poorly to ancient languages: generic multilingual encoders underperform, specialized language models yield anisotropic representation spaces, and labeled similarity data is unavailable. We study two fully unsupervised strategies - TSDAE and contrastive sentence embedding (CSE) - that adapt a specialized token-level language model into a corpus-specific sentence encoder using only raw sentences. On the philologically central case of biblical reuse in patristic literature (2,935 expert-verified parallels in Latin and Ancient Greek, from Augustine, Jerome, and Athanasius), we decompose reuse identification into two separately evaluated tasks-binary detection and correspondence retrieval-and benchmark the adapted encoders against multilingual, specialized, distilled, and supervised fine-tuned baselines, as well as on artificially noised data simulating HTR artifacts and scribal abbreviations. The adapted encoders outperform all baselines on both tasks, with complementary profiles: TSDAE leads detection given a large in-domain corpus, while CSE leads retrieval, reaches its optimum with as few as 4-8k raw in-domain sentences-a few tens of seconds of training on a laptop GPU-and transfers across works and authors, including to noisy post-ATR text when retrained directly on it. UMAP atlases relate the geometric effect of each strategy to the measured gains, and the full pipeline-segmentation, fine-tuning, cross-corpus semantic search-is made available to non-specialists through the online tool Paraphrasis.

In this paper, we present NEO, a unified framework providing language-guided NeRF editing for robotic manipulation. Our paper introduces (i) a language-guided object removal that combines neural field resampling with multiview-consistent progressive inpainting, (ii) a direct NeRF weight editing method utilizing knowledge distillation, composing original and edited NeRFs via a teacher-student model, enabling coherent modeling of future scene states before a robot executes an action, and (iii) the first benchmark (NEO-Dataset) for quantitatively evaluating NeRF scene editing methods suitable for robot manipulation. We show that our approach outperforms state-of-the-art baselines in scene editing tasks, including object removal and pick-and-place robotic experiments, yielding visually coherent and geometrically consistent edits that reduce artifacts commonly introduced by prior methods.

Articulated object manipulation requires an understanding of kinematic structure that is difficult and costly to learn from robot demonstrations alone. We introduce the Kinematic-Aware Articulation Interface (KAI), a structured intermediate representation that captures the kinematic structure of articulated objects. By embedding interpretable geometric and kinematic priors into policy learning, KAI provides a strong inductive bias aligned with the underlying structure of articulated motion. This design effectively improves sample efficiency, with gains particularly pronounced in low-data regimes: across six simulation tasks, our method achieves an average success rate of 82.9%, matching or surpassing baseline performance while using only half the demonstration data. Our method also exhibits robust generalization to unseen backgrounds and visual distractors, transferring from a single clean training environment to cluttered real-world scenes. KAI's action-agnostic design further enables co-training with human interaction videos to enhance real-world robustness: under diverse visual distractions, our method with video co-training achieves over 70% average success rate.

The goal of schema integration is, given a set of input schemata or tables, to derive a global, unified schema that is able to represent the concepts, attributes, and relationships of all input tables in a coherent fashion. This paper presents SINT-Flow, a schema integration framework composed of five LLM-based operators that can be combined into workflows to perform fully automated, end-to-end schema integration. In contrast to existing approaches, SINT-Flow can process denormalized source tables that contain attributes describing multiple entity types. During the schema integration process, these tables are decomposed into separate entity-specific relations. To evaluate SINT-Flow, we introduce SINT-Bench, a schema integration benchmark comprising 10 schema integration tasks consisting of altogether 93 relational tables, including tables that describe multiple types of entities. We evaluate SINT-Flow using GPT-5.2 as well as the open-weight model Qwen-3.6-27B as alternative backbone models. Using these models, SINT-Flow achieves F1 scores of at least 96% for entity-type detection, 85% for attribute detection, and 83% for schema mapping. Furthermore, we perform an ablation study to prove the utility of the applied self-consistency strategy as well as the inclusion of a review loop into the schema matching operator.

Learning the informative tactile representation while effectively adapting it to pretrained Vision-Language-Action (VLA) models remains challenging at both the data and modeling levels. At the data level, limited task-specific demonstrations constrain representation quality, whereas large-scale pretraining incurs substantial costs. At the modeling level, existing methods either focus on instantaneous contact states or model temporal interaction dynamics using 6D wrench sequences, leaving high-dimensional tactile signals underexplored. To address these challenges, we present τ, a touch-augmented VLA framework that learns an action-conditioned spatiotemporal tactile representation from future visual supervision inspired by the Joint-Embedding Predictive Architecture (JEPA), and fuses it with vision-language features for action generation under limited data. This supervision operates in latent space and is used only during training, adding no deployment overhead. We also introduce TacAura, a dataset of synchronized vision, proprioception, and vision-based tactile signals across four representative contact-rich manipulation tasks. Experiments show that τ outperforms existing models and generalizes to unseen objects and scenes, delivering improved manipulation performance and robustness

Real-world evaluation is a bottleneck in developing generalist robot manipulation policies. Each rollout requires physical hardware and an operator to set up, reset, and score it. We introduce ArmnetBench v0.1, a benchmark run on a fleet of low-cost SO-101 cells under light on-site supervision. v0.1 validates this arm farm end to end and compares 7 policies across 12 tasks with both single-arm and bimanual configurations. Each policy is trained or fine-tuned on 50 demonstrations per task; the benchmark contains 2,518 policy rollouts and 600 reference demonstrations. All 3,118 episodes carry a three-way label (successful, suboptimal, or failure). Policy rollouts are human-scored, while demonstrations are successful by construction. Beyond evaluation, its quality-labelled trajectories support downstream learning, from reward and predictive world models to policies trained on mixed-quality data. The leaderboard is an initial comparison under this shared budget. We release the 3,118 core episodes in LeRobot v3.0 and RoboMeter formats.

A central question in the in-context learning literature is whether transformers can organize episode-level adaptation around different inductive-bias families. We study this question in a controlled setting through latent algorithm routing: route-like behavior in which the solver-family preference changes with the latent data-generating regime while prompt form is held fixed, remains stable under nuisance perturbations, and is selectively influenced by targeted activation interventions without large losses in answer quality. We introduce ROUTEBENCH, a diagnostic benchmark whose regimes differentially favor global shrinkage, sparsity, robustness, and locality, operationalized by ridge-like, lasso-like, Huber-like, and kNN-like family representatives. Across dense decoder-only transformers trained from scratch at 44M-612M parameters, a 306M model closes 80.9 percent of the oracle-routing gap and achieves route F1 of 84.1. The effect remains substantial under natural-language renderings, shuffled supports, lexical paraphrases, and a unified four-way routing setting. Stronger adaptive alternatives, including an input-conditioned soft mixture and an unsupervised Gumbel router, narrow the gap but remain below the 306M and 612M models on route F1 and OOD performance. Probe controls and matched activation-patching controls further show that route-relevant internal directions are decodable and functionally involved in solver-family-consistent output behavior. These results provide controlled evidence that dense transformers trained on ROUTEBENCH can develop route-like internal variables, but they do not establish universal routing in pretrained language models or unrestricted natural-language reasoning.

Vision-language models (VLMs) deployed on consumer hardware must decide when to answer and when to defer, and that decision depends on having a confidence signal that tracks correctness. A practitioner with a fixed memory budget faces a choice between a small model at full precision, the same small model quantized, and a larger model quantized into the same footprint -- three configurations that push the confidence signal in opposing directions. We measure, on identical inputs, how model scale and 4-bit quantization affect two confidence signals in the Qwen2-VL family: the confidence a model states in natural language, and its own mean token probability over the answer it generates. Across 5,700 predictions spanning six realistic photographic degradations at three severities, we find that scale sharply improves the model's internal uncertainty signal (mean error-detection AUROC 0.80 to 0.98 from 2B to 7B) while its verbalized confidence stays weak and often at chance (mean 0.61 to 0.69): the gap between what the model knows and what it says widens rather than closes with size. We find that 4-bit quantization is nearly free for accuracy (-1.6 points) but expensive for the confidence signal (internal AUROC 0.95 to 0.80, and the verbalized-confidence parse rate collapses from 99% to 64%). For a fixed memory budget the recommendation is therefore to prefer a larger quantized model over a smaller full-precision one: 7B-4bit gives both the best accuracy and the best uncertainty signal (internal AUROC 0.98) of the three configurations that fit. We frame the results as selective-prediction operating points so they translate directly into a deployment recommendation, and we argue that error-detection AUROC, not calibration error, is the metric that exposes the difference between the two signals.

A recently proposed method was able to generate current references with minimum stator copper loss (SCL) and torque ripple over the full torque-speed range of multiphase nonsalient permanent-magnet synchronous machines (PMSMs) with nonsinusoidal back-electromotive force (back-EMF). However, it relied on large lookup tables (LUTs) generated offline, requiring several hours for generation and large memory for storage. Since on-chip memory in industrial digital signal processors (DSPs) is limited, this hinders practical implementation. This paper addresses this problem by analyzing how the LUTs of that method can be simplified and optimized, focusing on the example of a symmetrical six-phase PMSM drive. Breakpoint selection (BS), LUT construction (LC), and interpolation approaches are evaluated to optimize the tradeoff between LUT simplification and performance. It is shown that the required memory and LUT generation time can be greatly reduced while keeping nearly the same feasible torque-speed area and performance. Simulation results confirm negligible performance degradation.

Vertical farming is a promising approach for increasing food production in controlled environments while reducing land use, water consumption, and dependence on external climatic conditions. However, its large electricity demand, mainly driven by artificial lighting, still limits its energetic and economic sustainability. This study evaluates the techno-economic potential of an optical-fiber daylighting system designed to transport concentrated sunlight directly to the crop growing zones of a vertical farm. A validated transient vertical farm model was coupled with Tonatiuh ray-tracing simulations to assess a single-axis solar-tracked Fresnel concentration system. Five optical fiber typologies and three daylighting strategies were compared in terms of crop production, specific electrical energy consumption, and light cost, and benchmarked against light-pipe and rooftop photovoltaic solutions. The optical-fiber system achieved an annual useful-PAR delivery efficiency of 19-27%, depending on fiber typology. By distributing daylight across all rack shelves and limiting heat gains through a UV-IR filter, it reduced electricity consumption by 35-89%, compared with 12% for the light-pipe configuration. The daylight-only strategy achieved the lowest SEEC, equal to 1.81 kWh kg^-1, but caused a 52% crop productivity reduction. Hybrid LED-optical-fiber strategies avoided this penalty and achieved electricity savings of about 50%, with a minimum SEEC of 3.16 kWh kg^-1. However, current optical fiber costs make the system economically unattractive, requiring an estimated capital cost reduction of about 90% to become competitive. Rooftop photovoltaics covered 22% of annual electricity demand and achieved an 11-12-year payback, making PV the most economically viable rooftop solar option under current market conditions.

Long-term memory systems store what a user says in an external store and retrieve it when a related query arrives. This interface rests on an assumption so natural that it is rarely stated: a memory that is needed will resemble the query that needs it. World knowledge breaks the assumption. A tree-nut allergy should change the answer to a macaron request through their almond-flour ingredient, yet the two texts share no cue a retriever can see. We call this failure mode the implicit-association blind spot and introduce InMind, a 125-task, expert-verified benchmark spanning ten life domains, with 113 tasks grounded in citable public sources. Its paired controls separate three explanations that existing evaluations conflate: the fact was never stored, the model lacks the bridging knowledge, or the fact was stored and never surfaced. The verdict is clean. With the decisive memory placed in context, the backbone answers 84.0 percent of indirect queries; when the same memory must be retrieved, six vector, graph, and agentic memory systems reach at most 14.4 percent, even though they recall the same facts on demand at up to 100 percent. An embedding with eight times the dimensionality raises answer-blind target recall for every system yet leaves the gap essentially intact. A minimal diagnostic probe that keeps memory visible before the query arrives recovers most of the gap, locating the failure in the query-conditioned interface itself and pointing to routing, deciding which facts must stay visible, as the open problem InMind is built to score.

The aim of this article is to verify whether integrating large language models (LLMs) with the Retrieval-Augmented Generation (RAG) architecture enables their transformation from standalone generative models into components of cognitive computing infrastructure with enhanced epistemic reliability. The study proposes an architectural approach based on locally deployed LLMs operating in on-premises environments without high-end GPU accelerators and examines their applicability in supporting regulatory management processes requiring continuous analysis and interpretation of legal acts. The proposed solution combines local LLMs with external knowledge repositories, creating a hybrid cognitive architecture in which the language model performs semantic interpretation while the RAG layer provides controlled knowledge retrieval, contextualization, and traceability of information sources. The implementation was validated using the Ollama and LM Studio execution environments together with the Polish language models Bielik and PLLuM running on consumer-class hardware. The results demonstrate that augmenting LLMs with RAG significantly improves the factual consistency, domain specificity and normative precision of generated texts while reducing the risk of unsupported content generation. Furthermore, the study shows that integrating RAG introduces auditability, controlled knowledge management and dynamic updating of regulatory information without retraining the language model. The findings indicate that locally deployed LLMs enhanced with RAG should be regarded not merely as text generation tools but as semantic processing modules within cognitive computing infrastructures supporting regulatory compliance and organizational decision-making in environments characterized by high legal and informational volatility.

Large language model (LLM) agents inherit reactive failure modes: escalation under provocation, sycophantic drift under flattery, perseveration when stuck. These are failures of propensity, not capability; they concern what a model does under sustained pressure, which training-time alignment reduces but does not eliminate at runtime. This research led to the Gubernaut Cognitive Controller (GCC), a model-agnostic runtime control layer in a Nelson--Narens monitoring--control loop: an object level reads and writes text, while a deterministic meta level reads only the numeric telemetry {intensity, valence, repetition} and returns a regulating posture. Because the meta level ingests zero tokens, no injection channel to the controller exists by construction (an architectural property, not yet adversarially tested); the text-exposed arbiter's compliance is measured, not assumed. We evaluate the GCC with a pre-registered, generate-once/judge-many protocol across a 4x4 matrix of four frontier models (GPT-5.5, Claude Opus 4.8, Gemini 3.5 Flash, Grok 4.3), each serving as both a generator and a judge. The regulated arm is calmer in 13 of 16 cells at p<.05 and 15 of 16 by sign; the three sub-threshold cells, including a -0.04 null, all fall on the single near-saturated host. The effect survives a lineage-independent fourth judge family (xAI), strong evidence that it is no artifact of shared judge style. The clearest mechanism is the recovery signature: arousal that integrates under attack and then decays, valence-gated, on de-escalation, replicating across all four families. Transcripts and panels ship with SHA-256 provenance and are re-judgeable; five failure modes are pre-registered. No consciousness claims are made.

Common chunking strategies in Retrieval-Augmented Generation (RAG) systems often create redundant chunks. These redundant chunks make the vector database bigger and slow down retrieval. A common fix is cosine-similarity thresholding. This method reduces each chunk to a single vector, then compares vectors using a similarity score. But a single vector can lose the fine-grained, token-level detail needed to tell a true duplicate apart from a chunk that just shares the same topic. We propose Cross-Attention Calibrated Deduplication (CACD). CACD checks each new chunk against an in-memory pool of chunks already kept, using a cross-encoder instead of a single pooled vector. This keeps token-level detail all the way to the final comparison. CACD combines three parts: the cross-encoder comparison itself, a New Information Score (NIS) that measures how much of a chunk is not explained by a candidate already kept, and a majority vote across several candidates rather than a single best match. NIS is calculated from the attention entropy of the cross-encoder. We tested CACD against five existing filtering methods, nine chunking strategies, and 18 configurations, all on the full SQuAD 1.1 validation set. In our experiments, CACD removes 9.75% of chunks on average. This drop rate is close to other semantic-level methods, and much higher than exact-match filters, which barely remove anything. In these experiments, CACD also processes each configuration in 51.0 seconds on average, about 27% faster than the strongest baseline, NERExact (69.6s), and about 7x faster than cosine-similarity filtering (356.7s). These results come from a single dataset, so we present them as an early comparison, not a general claim. Code for the baseline evaluation and for CACD is available at https://github.com/lehuyphuong/rag_bench and https://github.com/lehuyphuong/cacd_dedup.

Marine life monitoring is limited by strict energy constraints, poor underwater connectivity, and the high cost of transmitting raw multimodal data from remote deployments. This paper proposes a low-consumption underwater monitoring architecture that combines always-on edge sensing with selective high-performance local reasoning. The system follows a hierarchical master--satellite design in which ultra-low-power MAX78000/MAX78002 microcontrollers continuously monitor visual and acoustic signals, while an NVIDIA Jetson Orin NX is activated only for scheduled processing, event-driven analysis, or researcher interaction. Once active, the Jetson executes a fully local multimodal pipeline for data ingestion, visual target extraction, embedding-based indexing, species identification, retrieval-augmented reasoning, and automated reporting. BioCLIP/OpenCLIP embeddings are used to organize mission data, marine taxonomic references, scientific documents, and operational metadata in local ChromaDB collections. A dedicated identification layer combines visual similarity search, centroid-based classification, and supervised classifiers to support adaptive species recognition. A LangChain-based multi-agent framework coordinates query routing, structured analysis, energy management, hardware reconfiguration, and report generation. The architecture is evaluated through visual and acoustic monitoring case studies. The proposed system bridges ultra-low-power continuous sensing with local multimodal intelligence, enabling underwater stations to produce structured, researcher-ready knowledge while compressing local data for flexible acoustic, optical, or satellite transmission, minimizing both energy use and communication overhead.

Self-improving agents accumulate capability by repeatedly rewriting procedural policies, controllers, or heuristic rules. They typically rely on self-authored tests or metrics to decide whether to accept subsequent edits. The agent controls both the optimized object and its verifier. As a result, self-assigned scores can remain near perfect while real deployment performance degrades or stays low. We study this problem through the verifier--deployment gap. This gap refers to the discrepancy between an agent's self-authored verification signal and a sealed deployment evaluation that the agent cannot observe or access. We ask how self-authored verification fails under iterative policy-and-test rewriting, how the failure changes with capability, and how little exogenous trust is sufficient to prevent real regressions from being deployed. To address this problem, we introduce a Sealed Exogenous Acceptance Loop (SEAL). SEAL retains self-authored tests but compares each candidate with the incumbent through a fixed harness-side audit. The agent cannot author or inspect the audit, receives only accept/reject, and the whole incumbent state is retained after a clear regression. Our experiments show that this problem often appears in heuristic learning settings. These settings require trial-and-error discovery of the target objective. We further find that failures of self-written verification are stratified by capability. Weaker agents tend to damage previously acquired strategies behind easy self-tests. Stronger agents are more stable, but they still mismeasure the deployment distribution. Standard self-written constraints do not reliably close this gap. In contrast, SEAL outperforms unprotected baselines across six models and three random seeds. Reliable self-improvement need not abandon self-verification, but it requires at least one deployment-acceptance signal outside the agent's control.

Diffusion Policies (DPs) are able to perform complex manipulation tasks. However, DPs are typically trained by minimizing a denoising objective, which provides limited control over generalization in the finite-data regimes common in robotics. In this letter, we propose PAC-DP, an approach that increases the performance of DPs in robotic manipulation tasks. By modeling the DP as a Bayesian neural network, and defining a PAC-Bayes generalization bound, we derive a novel training objective that augments the standard denoising loss with a Kullback-Leibler divergence regularizer between the posterior and prior parameter distributions. From the theoretical perspective, our approach provides a principled approach to regularize the training of DPs without significantly increasing the training time. From the practical point of view, experimental results demonstrate improved denoising performance, lower variational negative log-likelihood, and higher success rates across multiple robotic manipulation benchmarks. Crucially, the largest improvements are observed in low-data training regimes and complex tasks, establishing PAC-DP as a theoretically grounded framework for robot policy learning.

Large language models (LLMs) process text through subword tokenizers rather than directly reading characters or words. Because these tokenizers are trained predominantly on English-centric corpora, they introduce a systematic and often overlooked disadvantage for many non-English languages. In this work, we quantify this tokenizer tax for Indian languages using the FLORES-200 parallel corpus, measuring tokenization fertility across six widely used tokenizers and fourteen languages. Under cl100k_base (used by GPT-3.5 and GPT-4), Indian languages experience an average 8.0x tokenization tax relative to English, reaching 13.0x for Malayalam, reducing the effective context window to as little as 12% of that available to English users for equivalent semantic content. We identify the primary mechanism behind this disparity: failed byte-pair merges that leave text fragmented into single-byte tokens, with merge failure strongly correlating with tokenizer tax (Pearson r = 0.89). We further show that this phenomenon is not an inherent property of Indic scripts but a consequence of tokenizer design. Multilingual tokenizers such as XLM-R and OpenAI's o200k_base reduce the average Indic tokenizer tax by 73%, demonstrating that the disparity is largely remediable. Beyond token statistics, we quantify a practical consequence by showing that, under fixed context budgets, Indian-language documents preserve substantially less original content than equivalent English documents. Finally, we examine the relationship between tokenizer fertility and reading comprehension performance on the Belebele benchmark, finding that the apparent correlation is largely explained by language resource availability rather than tokenizer behavior alone.

Large Language Models (LLMs) have shown strong potential in financial reasoning, but existing benchmarks often evaluate domain knowledge, numerical reasoning, long-context understanding, and tool use in separate settings. This limits their ability to assess realistic professional workflows that require auditable, context-grounded, and tool-executable decisions. We introduce \textbf{INS-ActBench}, a comprehensive benchmark for evaluating professional actuarial capability in LLMs. INS-ActBench contains 12,050 Q\&A pairs from public exams and sample questions released by 16 actuarial associations. It covers three subsets: \textbf{INS-Act-Know} for standardized actuarial knowledge, \textbf{INS-Act-Case} for long-context insurance case reasoning, and \textbf{INS-Act-Practice} for spreadsheet and R-code tasks with verifiable numerical outputs. Experiments on nine representative LLMs and human actuarial experts reveal a clear capability boundary: frontier LLMs perform strongly on standardized knowledge, but remain much weaker in case reasoning, tool-based workflows, and jurisdiction-sensitive practice. INS-ActBench provides a reproducible foundation for developing actuarial LLMs toward reliable professional assistance. The code is available at https://github.com/FDU-INS/INS-ActBench.

Language-model benchmarks collapse two distinct measurement questions into a single accuracy score: whether a response reached an evaluable state, and whether its answer was judged correct. We introduce a two-layer evaluation framework that separates scorer-independent execution evidence, including termination, answer exposure, parseability, and completion length, from scorer-dependent correctness. Across 2,550 outputs from five fixed Qwen and DeepSeek configurations on MATH and ARC-Challenge, matched 2,048-token limits produce sharply different execution mixtures: 49 of 450 Qwen MATH outputs terminate without a final answer, compared with 5 of 300 DeepSeek MATH outputs and none of the 750 ARC outputs. Among the same 300 DeepSeek MATH question-model pairs, no missing-final length termination is observed at 8,192 tokens. A coverage-audited targeted verification study further shows that candidate-selection and aggregation policies can substantially alter comparative accuracy estimates. These results demonstrate that accuracy conflates execution case mix with verification policy. Evaluations of test-time methods should therefore report pre-intervention execution states, verification coverage, and scorer provenance alongside accuracy.

Humans plan physical interactions by imagining the possible outcomes of candidate actions. However, existing visual world models primarily capture appearance dynamics while overlooking the tactile states that govern contact-rich interactions, potentially producing imagined futures that appear visually plausible but violate physical dynamics. We introduce FeelWorld, a hierarchical visuo-tactile world model that jointly predicts future visual latents and three tactile states. FeelWorld organizes these states hierarchically as contact state, a 3D tactile latent that encodes force-related information, and slip state. These states are jointly predicted by a shared latent dynamics model with explicit supervision. To prevent irrelevant tactile signals during free-space motion from degrading visual prediction, we introduce a contact-gated asymmetric attention mechanism that maintains a visual-only prediction pathway before contact and enables joint visuo-tactile dynamics prediction during contact. The model is further trained with autoregressive rollouts and context noise injection to improve robustness to compounding errors. The predicted contact and slip states also support contact-aware CEM planning. Experiments on chip grasping, fruit grasping, and USB insertion show that FeelWorld reduces 10-step LPIPS from 0.084 to 0.058 and maintains an LPIPS that is 61% lower than that of the visual baseline after an 80-step autoregressive rollout. FeelWorld also achieves an average zero-shot planning success rate of 81.7%, providing an effective approach for incorporating tactile sensing into world models.

Standard depth sensors systematically fail on transparent surfaces, creating corrupted 3D maps and severe navigation hazards. While specialized hardware sensors can detect glass, they lack modularity and have extensive hardware dependencies. Consequently, learning-based monocular depth estimation has emerged as a compelling alternative. However, domain-specific glass-aware monocular depth estimators struggle with unfamiliar indoor layouts; restricted by the severe scarcity of real-world glass depth annotations, they fail to generalize zero-shot to new settings. This motivates us to explore whether the extensive priors of text-to-image diffusion models can enable generalizable perception of transparent surfaces. We introduce SILICA, a unified pipeline leveraging these priors to jointly predict glass segmentation and glass-aware depth. This mutual information exchange establishes a robust spatial hierarchy, entirely eliminating the need for paired real-world glass depth annotations. Subsequently, we use the predicted segmentation mask to explicitly filter incorrect glass depth points from standard sensors, recovering accurate metric glass depth for downstream 3D mapping and autonomous collision avoidance. Supported by our novel Mirage 18k dataset, extensive experiments demonstrate that SILICA achieves remarkable zero-shot transfer across diverse, unseen environments, outperforming state-of-the-art models by almost 20% and setting a new benchmark for transparent surface perception.

Repository-level code generation relies on heterogeneous evidence whose relevance, compatibility, and completeness are inherently uncertain. Similar-code examples, repository context, and project-specific APIs may provide complementary information, but can also introduce noisy, redundant, or conflicting signals. Existing retrieval-augmented approaches primarily optimize retrieval relevance without explicitly modeling how uncertainty in retrieved evidence affects downstream generation. We introduce OpenCoder, an uncertainty-aware framework that estimates source-specific uncertainty, uses it to filter and rank heterogeneous evidence, and guides generation, verification, and repair. A factorial analysis over API knowledge, repository context, and similar-code evidence reveals no universal additive source ranking; instead, significant cross-source interactions depend on the accompanying evidence and LLM backend. On an expanded 32-task RepoExec-inline evaluation, OpenCoder improves GPT selected-output correctness over Baseline RAG from 56.25\% to 78.13\%. However, it matches a verification-and-repair control, and the corresponding Gemini improvement is not statistically supported, indicating backend-dependent benefits. Target-aware API refinement also substantially improves API-set retrieval. These findings support treating uncertainty as an actionable control signal for repository-level retrieval, verification, and repair.

Relevance is a query-dependent estimate of whether a document or excerpt contains useful evidence. Existing retrieval agents use relevance to select top-$k$ content, but document relevance alone cannot localize, compose, or verify the evidence required by complex questions. Direct Corpus Interaction (DCI) enables such fine-grained operations through grep-style exploration, but its relevance-agnostic search can expose useful clues late and delay convergence. Recent advances use relevance to narrow the corpus into a working space for interaction. Once interaction begins, however, relevance still does not directly guide which documents grep searches first or distinguish informative excerpts from a broad set of matches to let LLMs see them first. We introduce the Relevance-Aware RipGrep Search Agent (RARG), which turns relevance into an execution prior for corpus interaction. RARG provides coarse-to-fine relevance guidance: it orders documents for sequential 'ripgrep' traversal to expose globally relevant clues earlier, initializes promising entry points with query-relevant paragraphs, and reranks grep matches to surface informative excerpts that document-level ranking may otherwise obscure. Across challenging browse question answering and reasoning-intensive retrieval, RARG improves the accuracy--efficiency frontier over retrieval-based and direct-interaction agents. These results demonstrate that relevance-aware interaction enables faster and more reliable search convergence.

Modern coding agents are usually evaluated by whether they eventually produce a correct patch, but patch generation depends on an earlier context-acquisition stage: finding the repository files needed for the task. We introduce Agent Retrieval Bench, a file-level benchmark for this upstream retrieval problem. Samples are built from real coding-workflow signals and evaluated against frozen base-commit repositories, with relevance defined by what an agent needs next rather than direct query-file semantic similarity. The benchmark covers four positive-retrieval tasks: code2test, comment2context, trace2code, and edit2ripple; a fifth subset evaluates selective retrieval using natural evidence-backed no-gold cases and counterfactual wrong-repository controls. Agent Retrieval Bench contains 427 samples across 25 repositories: 345 positive examples, 50 natural no-gold examples, and 32 counterfactual controls. The corpus includes 308 base-commit snapshots, 392,000 files, and 7.9 million chunks. We evaluate lexical retrieval, RepoMap, open-source embeddings, selective abstention, and logged agent context selection. No single retrieval family dominates: Qwen3-Embedding-4B has the best sample-weighted MRR on positive samples, Qwen3-Embedding-8B the best Recall@20, and RepoMap the best budgeted context yield at 8K tokens, with task-level winners differing substantially. Selective thresholds calibrated with counterfactual controls do not improve selective success on natural no-gold cases, revealing a calibration gap. Logged trajectories also miss every gold file on 27-35 percent of samples. A controlled seed-intervention pilot finds that retrieval-derived initial context yields higher file F1 with less post-seed exploration than random non-gold context, while oracle gold context shows substantial remaining headroom.

Mobile manipulation requires robots to identify Floor Affordance (FloAff) that maximizes downstream manipulation success rather than merely ensuring navigation feasibility. FloAff prediction is a target-conditioned local spatial reasoning problem, yet existing methods suffer from representation ambiguity caused by irrelevant spatial context and arbitrary object orientations, while entangling shared and task-specific knowledge across heterogeneous manipulation skills. To address these challenges, we propose a unified framework for FloAff prediction from egocentric multimodal perception, consisting of canonical representation learning and progressive affordance prior learning. Specifically, we introduce a Canonical Floor Affordance Representation (CFAR), which learns canonical interaction geometry by preserving affordance-relevant local structure while eliminating nuisance spatial variations unrelated to robot base placement. We further propose Progressive Floor Affordance Learning (PFAL), which learns transferable FloAff priors from a foundation manipulation task and progressively adapts them to heterogeneous downstream manipulation skills. To facilitate systematic evaluation, we establish the first cross-scene, multi-view FloAff-Kitchen benchmark covering diverse manipulation skills, scene layouts, furniture styles, and viewpoints. Extensive experiments on three benchmark settings demonstrate that our method consistently outperforms strong baselines, while ablation studies validate the contribution of each proposed component. Project page: https://csu-hero-lab.github.io/FloAff-Kitchen_Web/

The introduction of new technologies, such as surgical robots, is driving the vision of a connected, smart operating room (OR). However, realizing this vision requires a deep understanding of surgical workflows, which relies on realistic datasets capturing the actions of all OR personnel from both full room and surgical field perspectives. Acquiring such data in real ORs is prohibitively challenging due to factors such as ethics committee approvals, limited space for camera installation, and sterility regulations preventing the use of tracking markers. We present a step-by-step methodology for re-enacting complete surgical procedures in a reconstructed OR. This approach enables the creation of repeatable and annotatable workflow datasets for training activity recognition models, generating scene graphs, and formalizing surgical process models. Developed for robot-assisted ophthalmic surgery, our methodology combines expert consultation, structured workflow formalization, OR reconstruction, role-based training, real OR observation, and iterative recording with post-take debriefing. We provide concrete recommendations to allow other research groups to seamlessly adopt this methodology for their own surgical domains.

The increasing integration of renewable energy sources (RES) into modern power systems, particularly in weak grid environments, demands advanced solutions for voltage and frequency stability. This paper demonstrates that Enhanced STATCOMs (E-STATCOMs) with grid-forming capabilities outperform synchronous condensers (Syncons) in wind farm applications. Using a power-admittance-based linear modeling framework and validated through detailed EMT simulations, the study shows that E-STATCOMs provide superior damping of sub-synchronous resonance (SSR), faster dynamic response, and effective suppression of DC components, transient overvoltages, and new oscillatory modes introduced by series compensation. Unlike Syncons, which are limited by mechanical inertia, E-STATCOMs offer tunable control, virtual impedance emulation, and scalable fault current support. These advantages make E-STATCOMs a more robust and flexible solution for stabilizing renewable-dominated grids, especially those with long transmission lines and series compensation. The paper also aims to provide a benchmark E-STATCOM design criterion that enables stakeholders, such as transmission system operators (TSOs), to assess its suitability without the need for exhaustive EMT simulations.

Conversational stance detection has shifted from static text analysis to dynamic multimodal modeling. However, existing benchmarks exhibit three key limitations: failure to capture the dynamic evolution of beliefs, particularly during stance reversals; difficulty in disentangling affective states from logical reasoning; and neglect of the critical role of multimodal cues in resolving pragmatic ambiguities such as sarcasm. To address these limitations, we propose StanceFlip, a benchmark designed for multimodal conversational stance flipping forecasting over multi-turn dialogues across five modalities and multi-scenarios, which includes two novel subtasks: 1) Multimodal Stance Sextuple Extraction, extracting holder, target, emotion, sentiment, stance, and rationale as static state snapshots of dialogue to capture fine-grained cognitive structures. 2) Dynamic Stance Flip Attribution, tracking stance reversals across the conversation and identifying their underlying triggers. Alongside the dataset, we propose a dedicated framework, named ConStaFF, for Multimodal Conversational Stance Flipping Forecasting (MCSFF). Built upon a large language model, ConStaFF performs end-to-end stance reasoning, with a Thought-of-Stance (ToS) reasoning framework and a self-reflective verification mechanism integrated for structured stance modeling and faithful flip attribution. Specifically, ToS decomposes the reasoning process into specialized cognitive personas to formulate target propositions, resolve cross-modal conflicts, and infer historical stance trajectories. Extensive experiments show that our approach achieves state-of-the-art performance on both sextuple extraction and flip-trigger attribution, outperforming strong multimodal large language model baselines by substantial margins.

Social robots that rely on large language models for conversation are unable to retain information across sessions. This absence of memory violates social expectations, potentially preventing the formation of persistent relationships. This paper presents a lightweight episodic memory module that integrates vector-based semantic retrieval with an LLM-controlled dialog system, deployed on the humanoid robot head Kim. The module employs a hybrid scoring function combining cosine similarity with a memory strength metric to retrieve contextually relevant past interactions and inject them into the generation prompt. The system was evaluated in a within-subjects video-based online study (N = 43) using the Human-Robot Interaction Evaluation Scale (HRIES). Results show that episodic memory significantly increased perceived sociability (d = 0.60, p < .001), with the strongest effects on perceived trustworthiness (d = 0.62) and warmth (d = 0.56). Perceived disturbance remained unchanged (d = 0.00), indicating that the implemented approach to personalized recall did not trigger privacy-related discomfort or uncanny valley effects. These findings suggest that episodic memory serves as a social lubricant in embodied Human-Robot Interaction, enhancing relational quality without eliciting negative affective responses.

Generalizable robot manipulation requires policies that can anticipate how visual scenes evolve while executing language instructions. While recent Vision-Language-Action models benefit from large-scale pretraining, their predominantly static pretraining objectives provide limited supervision for physical dynamics and temporal causality, leaving control-relevant knowledge to be learned from downstream robot demonstrations. Video generative models offer a promising foundation by encoding rich spatiotemporal priors through future predictions. However, existing Video-Action Models either couple video and action prediction in a shared backbone, making policy adaptation harder to optimize, or under-utilize video information when guiding the action branch. In this work, we introduce DeVA, a Decoupled Video-Action model with specialized video and action experts, multi-level feature transfer, and physically salient guidance. DeVA transfers representations from multiple video layers to the action expert, enabling rich information exchange while making policy learning more tractable. It further supervises intermediate video features and the action stream with physically salient guidance (affordance/depth). Experiments on both simulation benchmarks and real-world deployment demonstrate strong performance with limited data, faster convergence than a unified architecture, and clear performance gains from physical guidance.

Previous research has shown that a human-like robot's acceptance heavily depends on the setting in which it operates and its ability to perform relevant tasks. This paper, first, reports on how our robot processes natural language to generate a multimodal, verbal response integrating emotional expressions based on an emotion simulation backend. Then, it describes how visitors were invited to speak with our robot in their own language at three different, public locations, where the robot was running continuously for several days. The TAM2 questionnaire results reveal that on average users were motivated to use the robot and found it rather useful and easy to use regardless of the specific location. However, public spaces like the tourist information and the city library seem to be a better fit for our interactive, robotic head than an office environment such as the building authority, where the willingness to interact was lower. Overall, the robot's multi-lingual responses were very much appreciated, but every fifth user found the response time too slow impeding the dialog flow, which remains to be improved in future work.

Reinforcement learning can produce robust humanoid controllers, but each new task is typically trained as a separate policy with its own reward design and training process. Motion imitation provides an alternative source of motor competence by training policies to track retargeted human motions, yet the resulting controllers remain reference trackers and are not directly usable as task policies. We propose a three-stage pipeline that turns motion-imitation skills into a reusable hybrid motion prior (HMP) for humanoid locomotion. First, an expert policy is trained to imitate retargeted human motion-capture clips. Second, the expert is distilled into a frozen architecture composed of a proprioceptive encoder, a residual vector-quantized (RVQ) codebook, and an action decoder. Third, task-level policies are trained to solve locomotion tasks by selecting discrete codebook entries while the HMP remains frozen. We evaluate the method on velocity tracking, point-goal navigation, and fall-recovery velocity tracking in simulation, and deploy the velocity-tracking policy on a real Unitree G1 robot. The distillation process preserves the tracking behavior of the expert, while the resulting HMP can be reused without retraining as the action interface for different downstream locomotion policies. The learned HMP reveals an interpretable codebook structure in which the number of active RVQ stages modulates the available gait patterns. We further show that training the codebook with the rotation trick improves latent organization and reduces downstream falls compared with a standard straight-through estimator.

Bridging the sim-to-real gap is a central problem in robotics, and the prevailing approach is to build increasingly accurate simulators. Here, we propose another approach based on renormalization: using effective, resolution-dependent parameters to absorb details omitted by the simulator and reproduce real behavior. These parameters may differ from measured physical values because they compensate for what the simulator leaves out. We demonstrate this mechanism analytically for proportional--derivative (PD) control at finite simulation frequency, where proportional feedback changes the effective derivative gain and derivative feedback changes the effective inertia. We then interpret dynamic rope manipulation and underwater swimming through the same perspective. Finally, we present a practical procedure for choosing observables, identifying omitted physics, and determining effective parameters. Renormalization offers robotics a complementary path across the sim-to-real gap: effective parameters, real behavior.

Low-thrust trajectory optimization is a core technology in deep-space mission design. Indirect methods based on Pontryagin's Minimum Principle (PMP) offer rigorous optimality guarantees, yet their practical application faces three bottlenecks: (1) transversality conditions must be derived case by case for each constraint type; (2) different dynamics models require repeated code rewrites; and (3) shooting equations are highly sensitive to initial guesses. This paper presents HELIOS (Heuristic Engine for Low-thrust Interplanetary Optimization System), a trajectory optimization agent built around a large language model (LLM). Given a physical problem described in natural language, the system autonomously performs PMP symbolic derivation, SymPy verification, C++ shooting-code generation, and numerical solution without human intervention. Key innovations include: (1) a constraint-adaptive derivation framework that unifies arbitrary constraints into psi(x,p)=0 form and automatically generates stationarity conditions for free parameters (e.g., gravity-assist turning angle); (2) dynamics-adaptive four-module code generation supporting non-standard dynamics (solar sail, J2 perturbation) without modifying the underlying template; and (3) a general derivation rule set covering critical error-prone points in PMP derivation. Experiments on 11 progressive test scenarios show that HELIOS correctly derives and solves problems from simple rendezvous (8 variables) to multi-leg stay transfers (48 variables), gravity-assist trajectories (17 variables), and solar-sail minimum-time transfers (8 variables). The best compilation success rate reaches 100% (11/11). A multi-model comparison (8 open-source LLM backends, total scores 250-905) verifies the model-agnostic architecture and reveals a positive correlation between model scale and derivation capability.

While reinforcement learning for legged robots has achieved high motor performance, it has been constrained by the limited exploration capability of actions confined to the joint space. To address this issue, this study proposes a new method, Wrench-Augmented Reinforcement Learning (WARL), which introduces a wrenche (force and torque) into the action space. The proposed method combines wrench-guided exploration with a success rate-based curriculum mechanism to expand exploration capabilities in the early stages of learning, with the ultimate goal of acquiring behaviors based solely on joint control. Experiments using a quadruped robot demonstrated that WARL can learn robustly across diverse terrains and motor tasks without requiring terrain-specific reward adjustments or complex curriculum designs. Furthermore, an ablation study verified the effectiveness of the Switching Curriculum, which gradually eliminates the wrench. On the other hand, we also show that introducing a wrench can encourage behaviors that do not sufficiently exploit the robot's physical embodiment. These findings suggest that while wrench-based exploration enhancement is effective for improving learning efficiency, designing it in a way that is consistent with the robot's physical structure is a critical future challenge.

Precise control of soft manipulators remains challenging due to the difficulty of developing accurate yet computationally tractable models for model-based estimation and control. Reduced Cosserat-rod models provide a physics-based and control-oriented description of soft-robot dynamics, offering an explicit alternative to purely data-driven input-output representations. In this paper, we propose a moving-horizon estimation (MHE) and nonlinear model predictive control (NMPC) framework for cable-driven soft manipulators based on reduced Cosserat dynamics. A smooth cable-length-driven modeling formulation is developed by approximating the complementarity relationship between cable tension and cable slackness, enabling cable-length control without direct tension sensing. Based on this formulation, an MHE method is introduced to estimate the reduced state and reconstruct the manipulator configuration from end-effector pose measurements and cable-length information. An NMPC controller is then formulated to achieve task-space control under cable-length and cable-rate constraints. The proposed framework is validated through numerical simulations and experiments. Simulation results demonstrate the effectiveness of the estimator and controller for pose and strain-related regulation on a multi-cable soft manipulator. Experimental results on a four-cable prototype further show that the proposed MHE-NMPC scheme can be implemented in real time and enables accurate end-effector position tracking through cable-length control.

Real-time deployment of Vision-Language-Action (VLA) policies necessitates asynchronous execution, wherein subsequent action chunks are computed concurrently with the execution of the current chunk, leading to prediction-execution misalignment and manifesting as inter-chunk discontinuities. Existing methods either superficially smooth chunk boundaries, require costly policy optimization, or exclusively forward-predict proprioceptive states yet neglect critical visual observations. In this paper, we propose \textbf{FutureRTC}, a plug-and-play adaptation framework that predicts execution-time observations and states for asynchronous VLA control without modifying the underlying policy. Specifically, FutureRTC features a state correction module to compensate for the discrepancy between rolled-forward and actual execution-time proprioceptive states and an observation prediction module that forecasts execution-time visual representations by leveraging robot motion as an explicit physical prior through motion-aware feature transport and reconstruction. Furthermore, we introduce a policy consistency loss to align the action chunks generated from predicted contexts with those produced under the expected execution-time inputs of the VLA policy. Extensive experiments across simulated and real-world environments demonstrate that FutureRTC achieves superior robustness to inference delays, resulting in smoother trajectories, faster execution, and consistently higher task success rates.

World Action Models (WAMs) have emerged as a powerful paradigm for embodied intelligence, yet the prevailing reliance on pixel-level video generation creates a fundamental bottleneck. Forcing models to reconstruct task-irrelevant visual details dissipates representational capacity and renders policies vulnerable to visual distractors. In this paper, we propose LeapBot-WA, which establishes a novel Predictive-Latent paradigm for WAMs by operationalizing the Joint-Embedding Predictive Architecture (JEPA) as a World-Anchor. Departing from the traditional reliance on visual synthesis, LeapBot-WA shifts the core of world modeling to Predictive Semantic Alignment, extracting abstract physical dynamics directly within a latent foundation space. To bridge the modality gap between non-Gaussian predictive features and diffusion priors, we introduce the Isotropic Semantic Autoencoder (ISAE), which reshapes the anchor's latent space into a diffusion-friendly manifold to prevent off-manifold drift. Furthermore, we design an Asymmetric Mixture-of-Transformers (MoT) architecture. During training, an Anchor Diffusion Transformer acts as a privileged dynamics expert to guide the Action Diffusion Transformer; at inference, this heavy dynamics branch is pruned, enabling zero-overhead execution. LeapBot-WA achieves state-of-the-art performance among predictive models on LIBERO and matches top-tier generative WAMs on RoboTwin 2.0 without requiring large-scale trajectory pre-training. It further demonstrates superior zero-shot robustness to unseen environments and successful real-world transfer, establishing a highly efficient and robust latent-centric paradigm for scalable robotic control. Code: https://github.com/LeapWM/leapbot-wa.

Operating constrained dynamical systems requires controllers to efficiently solve complex tasks while enforcing recursive feasibility and safety constraints. To address these competing requirements, we present Feasible Action for Optimal Control (FAOC), a novel control framework integrating Reinforcement Learning (RL) and Optimal Control (OC). The key contribution is a computationally efficient, optimization-based mapping algorithm that transforms the RL agent's action from a static abstract set into a state-dependent feasible parameter set of the Optimal Control Problem (OCP), guaranteeing strict satisfaction of the dynamical system's constraints. Thus, FAOC effectively combines the predictable safety of OC with the flexibility of RL. In contrast to prior work, the abstract action space of the RL agent does not require expert or heuristic design, and the OCP formulation is not compromised by the inability of RL to guarantee feasibility. We apply our approach to real-time motion planning for robot table tennis, which encapsulates these challenges. Via simulated experiments, we show that FAOC outperforms state-of-the-art baselines in both sample efficiency and closed-loop performance.

Cooperative perception through vehicle-to-everything (V2X) communication can overcome the inherent physical limitations of individual autonomous vehicles, such as occlusions and limited sensor range. However, the development of robust V2X algorithms, particularly those relying on unified spatial representations like bird's-eye view (BEV) representation, is hampered by the lack of large-scale, multi-modal, multi-task datasets. Moreover, collecting and annotating a large set of synchronized, real-world multi-agent data is prohibitively expensive. This has resulted in a landscape where existing V2X datasets are notably limited in both size and scope. To overcome this, we introduce SimBEV2X, an advanced synthetic data generation tool built on the CARLA simulator. SimBEV2X automatically creates randomized driving scenarios to collect multi-modal sensor data alongside various types of ground truth including 3D bounding boxes with unique track IDs, HD map information, BEV segmentation maps, and semantic occupancy voxel grids from both vehicles and RSUs. We also present the SimBEV2X dataset, the largest V2X perception dataset to date. The dataset comprises 258 scenes, each involving up to 8 connected vehicles and up to 4 RSUs across a variety of road networks. The SimBEV2X dataset is an order of magnitude larger than existing V2X datasets and contains 102,200 frames, 588,520 lidar point clouds, more than 3 million images, over 27 million bounding boxes, and a comprehensive set of other annotations. Finally, we establish a strong baseline on the SimBEV2X dataset using CoopDet3D and propose CoBEVFusion, a novel architecture that combines CoopDet3D with fused axial attention (FAX) for context-aware multi-agent feature aggregation, resulting in superior performance. SimBEV2X, the SimBEV2X dataset, and CoBEVFusion are available at https://simbev2x.org and https://github.com/GoodarzMehr/SimBEV2X.

Many recent robot policies pursue stronger control by using large pretrained vision-language models (VLMs) as the action backbone. We introduce WorldDiT, a unified diffusion transformer architecture that couples action generation with visual world modeling and achieves strong performance without a large pretrained VLM action backbone. During training, a single diffusion transformer generates continuous action chunks and predicts normalized RGB patch targets from future camera frames. Across four LIBERO simulation suites, WorldDiT lies on the reported Pareto frontier for total model parameters and mean success among methods reporting all four suites. These results provide a strong sub-billion-parameter baseline for future scaling studies.

Human-Robot Collaboration (HRC) in shared physical spaces using Augmented Reality (AR) interfaces is powered by Simultaneous Localization and Mapping (SLAM). Existing multi-agent SLAM systems rely on an edge server to combine visual findings of multiple resource-constrained agents, perform computation, and schedule updates to their local maps. However, the edge treats all agents uniformly and ignores the fundamentally different latency requirements of heterogeneous HRC agents: robots and head-mounted AR users. This uniform resource allocation often results in high lag for user manipulation, as it does not meet the stringent latency requirements of AR. In this work, we design, implement, and evaluate SHARE, a user-centric SLAM system that strategically prioritizes AR user experience while maintaining accurate tracking performance for robots. SHARE builds a first-of-its-kind experience model for HRC agents and adaptively adjusts transmission priorities to match it. To reduce end-to-end latency, SHARE leverages the redundancy of visual features acquired by agents in shared human-robot workspaces to reduce computation time induced by edge-based processing. Real-world deployment with commercial AR headsets and a ground robot achieves 13.22 ms average latency for AR users (43.3% reduction from baseline) while maintaining sub-2-centimeter tracking accuracy. User studies further reveal statistically significant improvements in user perception.

This exploratory study examines whether a large multimodal language model, GPT-5.1, can serve as the high-level controller of a physical mobile robot despite having no prior embodiment, no training in simulated environments, and no exposure to sensorimotor experience. Using only low-resolution first-person images and a discrete action set, the model was tasked with navigation and object-directed behaviors such as locating and contacting a target toy. Across multiple trials, GPT-5.1 demonstrated emergent capabilities that suggest elements of spatial reasoning and physical understanding. These included maintaining short-term memory of object locations after they left the camera frame, inferring the physical consequences of its own movements, and executing coherent action sequences such as colliding with an object and reversing to visually verify the outcome. At the same time, the model displayed inefficiencies and perceptual limitations, including imprecise alignment strategies and occasional misidentification of distant distractors. Overall, the results indicate that GPT-5.1 exhibits signs of world-model-like behavior in an embodied setting, despite the absence of any embodiment-related training, a finding that challenges long-standing views in cognitive science and robotics which hold that a physical body is a necessary prerequisite for developing such forms of intelligence. The findings motivate deeper investigation into the emergence, limits, and robustness of physical understanding in large language models.

Autonomous laboratories automate experimental execution, but a campaign must also decide which recovery pathway merits optimization. We formulate this as a sequential decision problem with a discrete pathway-identification stage and a continuous within-pathway optimization stage under heterogeneous experimental costs. Our implementation, Coactive learning, combines a cost-sensitive Bayesian hypothesis-discrimination policy motivated by EC2 (Golovin et al., 2010) with Gaussian-process Bayesian optimization (Srinivas et al., 2010). Under explicitly stated assumptions, the expected spend of one fixed-budget campaign attempt is bounded by the expected pathway-identification cost plus the capped within-pathway optimization budget. We evaluate the method on synthetic benchmarks constrained by selected results reported for PNNL's CICERO selective-precipitation study (Ritchhart et al., 2026). The method performs comparably to an oracle-pathway Bayesian-optimization reference and to a strong split-plate baseline that discriminates pathways with its first plate, without receiving an oracle label for the correct pathway. It is given a candidate hypothesis space and a diagnostic likelihood model. On an NdFeB-inspired instance, it avoids the simulated penalty of a commit-first baseline that initially selects a plausible but inferior hydroxide pathway. This hypothetical wrong-first-commitment scenario is motivated by the hydroxide-oxalate performance contrast reported by CICERO. We characterize the sensitivity of these conclusions to the assumed cost model. The code and benchmark are open source.

A robot carrying a persistent, behavior-annotated map faces two planning questions, and its memory answers only one well. The \emph{spatial-navigation} question -- how to walk around a room -- we address first and report a negative: building on Vision--Language--Motion Maps (VLMM), a behavior-aware planner cost cuts a planning-time objective by $\sim$35\% over 28 AI2-THOR scenes, but under closed-loop execution the real benefit nearly vanishes ($\sim$4\%) and an on-demand vision--language model (VLM) does as well. The \emph{resource-allocation} question differs: under a limited perception budget, what should the robot re-observe now to keep its map fresh? Framing re-perception as this attention decision, we show a persistent map's memory (change-history, or even just recency of last sighting) yields the best schedule (held-out), matching an oracle, while the memoryless VLM prior is poor. Because the schedule reallocates budget toward what matters, memory's benefit concentrates on the important objects ($\sim$1.6$\times$ the mean), and a downstream fetch task confirms fewer wasted trips; the gain grows with per-instance heterogeneity exactly as a Cauchy--Schwarz bound predicts -- it equals $\mathrm{Var}(\sqrtλ)$, the variance of root-volatility. With a real CLIP prior on rendered objects the advantage is $+21$--$26\%$. The map's distinctive value appears when the task is \emph{language-conditioned}: told what to track, VLMM grounds the relevant objects (open-vocabulary) and tracks their change (memory), beating even a strong relevance-weighted recency baseline ($+2.5\%$) -- so its motion channel adds value beyond a last-seen timestamp -- and an on-demand VLM ($+8.9\%$); neither language nor dynamics alone suffices. The map earns its keep not by telling the robot how to walk around a room, but by telling it what to pay attention to.

Rapid preliminary design of low-thrust interplanetary rendezvous trajectories requires parameterizations that combine computational efficiency, flexibility, and sufficient solution quality. This work presents a shape-based method in which the heliocentric cylindrical coordinates are parameterized using clamped B-splines. The differential flatness of the low-thrust cylindrical dynamics is exploited to recover the control acceleration algebraically from the shaped trajectory and its derivatives. The clamped B-spline parameterization satisfies the endpoint position and velocity conditions analytically, reducing the original fixed-time, minimum delta-v problem to a finite-dimensional nonlinear program in the free interior B-spline control points. The formulation therefore requires no numerical propagation of the equations of motion; numerical quadrature is used only to evaluate the objective function. Numerical campaigns are conducted for low-thrust rendezvous transfers from Earth to Mars, Mercury, the near-Earth asteroid 1989 ML, and comet Tempel 1, and the results are compared with high-order hodographic shaping. Across the tested fine-grid campaigns and among the finite solutions retained, all B-spline configurations reduce the mean, median, and minimum accumulated velocity increment relative to the hodographic benchmark. In particular, a ten control-point quintic B-spline is identified as a favorable compromise between solution quality and computational efficiency, while higher-dimensional parameterizations are effective for refining selected transfer opportunities.

While vision-language-action models have demonstrated impressive zero-shot manipulation capabilities, they remain fundamentally black box policies that are difficult to interpret, adapt, or correct when they inevitably fail. In this work, we propose ARCHITECT, a framework that treats robot policy acquisition as an interactive program synthesis task. ARCHITECT leverages the reasoning capabilities of LLM coding agents to synthesize modular robot programs that utilize a suite of perception and control tools. Unlike end-to-end models where distribution shift leads to unpredictable, cascading failures, our modular architecture allows users to isolate failures and localize feedback at the level of abstraction required. We introduce an iterative process where a human supervisor provides natural language corrections to steer the policy. These corrections are grounded in the policy code by program execution traces and distilled into a persistent skill library, a form of long-term in-context learning which enables the agent to accumulate a repertoire of reusable, interpretable behaviors. In a benchmark evaluation on a Franka Panda robot, ARCHITECT outperforms state-of-the-art VLA models and program synthesis baselines on complex, long-horizon tasks, including articulated object manipulation and cloth folding. Our results demonstrate that the synthesized skill library enables the system to transfer to novel tasks with decreasing human intervention, providing a steerable and data-efficient alternative to black-box robot learning. Website: https://robo-architect.github.io/

We present $N_0$-TWAM, a tactile-native world-action model for contact-rich manipulation that predicts both future vision and future contact. To our knowledge, it is the first tactile world-action model trained at large scale, and it shows strong capability on contact-rich tasks. We pre-train $N_0$-TWAM at large scale with visuo-tactile joint training over tactile-rich demonstrations spanning six embodiments and 450 tasks. We use NeoForce, a unified force-based tactile representation, to form a physically grounded contact signal that conditions action generation. To improve long-horizon and multi-stage manipulation, we introduce tactile contact events for task staging and advance through them during execution. For real-time efficiency, we adopt an asymmetric Mixture-of-Transformers architecture that pairs a full-width expert for video prediction with slim experts for downstream action and tactile prediction. Evaluations on both real and simulated benchmarks justify the capabilities of $N_0$-TWAM across a range of contact-rich tasks, and demonstrate the benefit of data scaling for precise tactile and action prediction. In summary, $N_0$-TWAM endows a world-action model with predictive capabilities to foresee vision, touch and action, building a solid foundation for fine-grained manipulation on open contact-rich tasks. The codebase and model checkpoints will be made publicly available to foster further research and development in tactile-enabled robotic manipulation.

We present $N_0$-VTLA, a vision-tactile-language-action (VTLA) foundation model capable of (1) fine-grained contact-rich manipulation with tactile perception and tactile-feedback control, and (2) offline policy improvement from stored deployment data. Building on current vision-based backbones, we propose a training recipe for tactile integration consisting of visuo-tactile pre-training, staged tactile-pathway integration, and advantage-conditioned offline policy improvement. During pre-training, the policy learns broad contact priors from NeoData, our large-scale visuo-tactile robot dataset; to our knowledge, $N_0$-VTLA is the first VTLA model pretrained on tactile data at scale. During post-training, we augment the policy with a predictive tactile pathway that distills the contact patterns learned at scale into the fine motion adjustments required by downstream tactile-centric manipulation. For offline policy improvement, we introduce ALTER, an advantage-conditioned offline reinforcement learning method that converts relative progress and trajectory-event comparisons into binary advantage labels for policy training on a fixed deployment corpus, further improving task-specific learning on contact-rich skills such as deformable object manipulation. Across contact-rich benchmarks, $N_0$-VTLA outperforms strong baselines by wide margins: it wins all nine real-robot NeoReal tasks and reaches 63.8% mean success on a twenty-task simulation suite, against 44.0% for the strongest baseline. $N_0$-VTLA policies trained with ALTER reach 75-95% success on three long-horizon real-robot tasks. These results lay a foundation for versatile tactile-driven manipulation policies.

Phase-change memory (PCM) is a mature technology for fast, scalable, non-volatile data storage, with applications spanning embedded memory, as well as in-memory and neuromorphic computing. PCM predominantly relies on chalcogenide alloys, with $\mathrm{Ge_2Sb_2Te_5}$ (GST) as the industry standard. Yet in these alloys, the individual Ge, Sb, and Te atoms redistribute upon cycling, causing stochastic operation and ultimately device failure. To address this issue, elemental antimony was proposed as a PCM material, but it exhibits a metastable amorphous state that prevents reliable data retention. Moreover, tellurium and antimony can contaminate complementary metal-oxide-semiconductor (CMOS) production lines or act as unintended dopants, restricting manufacturing of PCM to dedicated fabs. Here we introduce elemental germanium (Ge) as a CMOS-native phase-change material that overcomes these fundamental limitations. In a vertical PCM cell architecture, Ge enables sub-nanosecond crystallization (240 ps, 40 times faster than GST), non-volatile data storage with excellent thermal stability ($>$110 °C for 10 years vs. $\sim$87 °C for GST), and a resistance drift coefficient approximately 60% lower than in GST. These results establish pure Ge, a standard semiconductor, as an alternative to chalcogenide phase-change materials, achieving superior performance in key metrics and enabling phase-change memory to be fabricated in standard semiconductor facilities.

The deployment of embodied agents in self-driving laboratories could accelerate scientific discovery, yet their reliability is constrained by the irreversible and safety-critical nature of chemical experiments. Progress is further hindered by scarce failure data and the lack of fine-grained evaluation protocols. To address these challenges, we introduce LabRobFail, a failure-centric framework for learning and evaluating robotic failure analysis in chemical laboratories. LabRobFail-Sim injects controllable failures at the control, physics, and semantic levels, enabling the construction of LabRobFail-Data, which contains over 20,000 trajectories across 70+ task scenarios, five failure categories, and 11 fine-grained failure types. LabRobFail-Bench evaluates six capabilities spanning task understanding, failure detection, temporal localization, severity assessment, failure classification, and actionable correction. We further develop LabRobFail-VLM, a domain-specialized vision-language model that generates structured failure diagnoses and recovery instructions. On seen environments, it achieves 92.58% failure-detection accuracy and 85.58% temporal-localization accuracy, substantially outperforming general-purpose VLMs. When integrated as a real-time supervisor, it improves downstream VLA task success rates by 10-20 percentage points, demonstrating the value of fine-grained failure understanding for closed-loop recovery and reliable laboratory autonomy. Our code and data are available at https://github.com/Su-ISE-2001/SciRobo

Manipulating objects with hidden internal state, such as a latched microwave, forces a robot to probe before it can act. Yet a robot that has solved an instance once re-runs the same probes whenever it encounters that instance again, because existing cross-episode memories target task success and organize reuse around states, not the object or the cost of re-exploring it. We present Instance-Oriented Memory (IOM), an object-centric framework that amortizes this exploration: from a single encounter that uncovers the hidden state, whether or not it succeeds, IOM records a short procedure for manipulating that instance, keys it on the object's identifiable features, and injects it as a soft bias on a procedure-conditioned policy. A later encounter recognizes the object and recalls its procedure instead of re-exploring. We instantiate this distillation with an off-the-shelf vision-language model (VLM) that parses each encounter into the procedure without task-specific training. Across four articulated-object tasks, two in simulation (microwave, door) and two on a real robot (bottle, cabinet), an oracle procedure memory cuts manipulation operations by 16-30% over re-exploration at non-regressing success, and the VLM instantiation recovers 69-88% of that saving out of the box. Because the procedure is a soft bias on a feedback-driven policy, an incorrect memory is recovered from rather than obeyed: success holds even when a retrieved procedure is wrong, as for $\approx$12% of door instances. Across all tasks the benefit is purely one of efficiency: success never regresses, and on the real robot even improves. Code will be released upon acceptance.

Conventional frame-based imaging for active stereo systems has encountered major challenges in fast-motion scenarios. However, how to design a novel paradigm for ultrafast depth sensing remains an open issue. In this paper, we propose a novel problem setting, namely active event-based stereo vision, which attempts to integrate binocular event cameras and an infrared 2D pattern projector for high-speed dense depth sensing. Technically, we first build a stereo camera prototype system and present a real-world dataset with over 21.5k spatiotemporal synchronized labels at 15 Hz, while also establishing a realistic synthetic dataset with stereo event streams and 23.8k synchronized labels at 20 Hz. Then, we propose ActiveEventNet+, a lightweight yet effective event-based stereo matching neural network that learns to generate high-quality dense disparity maps from stereo event streams with low latency. Our ActiveEventNet+ mainly involves three innovations: incorporating lightweight blocks into event-based stereo matching frameworks, designing a novel cost volume with dynamic interactions between stereo pairs, and presenting an effective temporal consistency architecture to fully use rich temporal cues in event streams. The results show that our ActiveEventNet+ outperforms state-of-the-art methods while significantly reducing computational complexity. Our solution offers superior depth sensing performance compared to conventional frame-based stereo cameras in high-speed scenes. In particular, the lightweight ActiveEventNet enables the prototype system to achieve real-time processing at speeds up to 150 FPS. We believe that this novel active event-based stereo vision paradigm can provide new insights into the design of future high-speed depth sensing camera systems. Our dataset and code can be available at https://github.com/jianing-li/active_event_based_stereo.

Robust object 6D pose tracking is critical for robotic systems operating in dynamic and occluded scenes. Per-frame estimators are accurate but computationally expensive, while current trackers struggle with fast motion and complete occlusion due to their reliance on continuous visibility. To address these challenges, we present RRTrack, an efficient, recoverable object 6D pose tracker that enables robust tracking through fast motion and target disappearance--reappearance. RRTrack introduces a 2D--6D closed-loop tracking strategy that integrates memory-based video object segmentation (VOS) with 6D pose refinement. The 2D branch maintains target localization, and the 6D branch verifies geometric consistency before memory updates. In addition, a DINOv2-based dual-bank template matching module is developed to recover lost targets by jointly exploiting offline synthetic templates and online observation anchors while maintaining real-time efficiency. We also introduce a synthetic RGB-D benchmark comprising three robotic scenarios with fast motion and full occlusion. Experimental results on the synthetic benchmark demonstrate that RRTrack improves equal-subset mean ADD-S AR by 66.3\% and ADD-S AUC by 65.7\% over FoundationPose while achieving 55.2 FPS. Real-world experiments further validate the robustness of RRTrack under noisy sensing conditions. Project page: https://github.com/7kevin24/RRTrack

Approaching a target position and holding station in flowing water is a fundamental and critical capability for robotic fish operating in natural aquatic environments. Despite decades of advances in enhancing swimming efficiency and maneuverability, this capability remains underdeveloped, largely owing to the insufficiently characterized, highly nonlinear fluid-structure interactions inherent to freely swimming robotic fish in flows. To bridge this gap, we propose the SWiFT framework, a Swimming With Flow Toolbox that enables the efficient exploration of an egocentric station-holding policy for a body and/or caudal fin (BCF) robotic fish in unknown and turbulent background flows via reinforcement learning (RL). Our SWiFT integrates a free-swimming flow-tank experimental setup with a highly efficient, physically consistent computational fluid dynamics (CFD)-based simulator and a systematic sim-to-real transfer pipeline. The resulting policy achieves substantial improvements over state-of-the-art methods across all metrics, most notably root-mean-square error (RMSE) of distance. Furthermore, we validated that egocentric feedback alone, without any explicit flow sensing, enables station-holding in unknown turbulent flows, closely mirroring the biological phenomenon of rheotaxis. Accordingly, the success of this egocentric station-holding policy not only advances robotic fish control toward real-world deployment, but also highlights SWiFT's promise as a foundation for tackling complex swimming tasks for underwater robots.

Controllers based on sampling and latent world models assign a predicted terminal cost to each candidate action sequence, choose the minimum, execute its first action block, and replan. This rule can fail even when the terminal cost perfectly and accurately reflects the true task objective in the physical world. Residual prediction error can give an infeasible sequence an anomalously low cost, and a larger proposal pool gives such errors more chances to outrank feasible alternatives. We call this conditional failure proposal overgeneration. In Cube candidate execution audits, increasing the total proposal budget from 72 to 288 reduces the feasibility of selection by minimum latent cost from .375 to .062 for position targets and from .344 to .031 for targets defined by position and yaw, although every larger pool contains a feasible sequence. We introduce Adjacent Set Action Reconstruction (ASAR). Among proposals with low cost, ASAR measures density from standardized early action prefixes and reconstructs a full sequence from an adjacent set with a light anchor from the sequence with minimum cost. On a Carry and Release evaluation set of 75 queries, Kernel ASAR improves event completion success over matching selection by 28.0, 24.0, and 18.7 percentage points under latent cost and by 18.7, 20.0, and 17.3 points under a trajectory reachability cost at 72, 144, and 288 proposals. Analysis of finite proposal pools characterizes selection risk from the lower tail, separation by a related radius support statistic, and sequence containment under an explicit local feasibility condition.

Safe quadrotor navigation in cluttered and dynamic environments depends not only on instantaneous geometric perception, but more critically on anticipating collision risks induced by relative motion. Conventional modular pipelines frequently suffer from perception latency, while end-to-end learning methods relying on implicit scalar rewards often struggle to extract reliable spatio-temporal features without physics-grounded supervision. To address this, we propose an anticipatory risk-guided reinforcement learning framework. Leveraging privileged simulator states, we construct a directionally aligned future collision risk map based on the Closest Point of Approach (CPA). Through an asymmetric actor-critic architecture, the network is trained to self-predict this structured risk, which explicitly guides the visual policy during deployment. A lightweight spatio-temporal encoder extracts motion cues directly from onboard depth sequences, bypassing explicit object tracking or optical flow estimation. Extensive simulated and real-world experiments demonstrate that our method effectively improves safety margins and flight efficiency in dense dynamic clutters compared to existing baselines. Furthermore, the learned policy achieves robust zero-shot Sim-to-Real transfer on a physical quadrotor, relying purely on abstracted spatio-temporal depth sequences and its self-predicted risk priors, validating the effectiveness of our approach and its robust generalization from simulation to reality.

Long-horizon manipulation tasks pose significant challenges for reinforcement learning due to sparse reward signals and long horizons. Automatic curriculum learning (ACL) has been proposed to tackle these challenges by progressively training agents on a sequence of tasks, from easier to more difficult. However, the success of ACL depends heavily on task-dependent specifications-such as well-defined task parameter spaces and difficulty measures-which are often manually crafted and difficult to generalize across diverse tasks. Recent advances in large language models (LLMs) offer a promising alternative by enabling the decomposition of complex tasks into meaningful subtasks using the LLMs' web-scale common-sense knowledge. This decomposition can provide a natural curriculum structure for efficient learning of long-horizon tasks. However, existing LLM-based methods typically rely on hand-designed dense reward functions to learn each subtask, which can introduce bias and still requires significant human supervision. In this work, we propose LLM-enhanced automatic curriculum learning (LEACL), a framework that integrates LLMs and ACL to address these limitations. Specifically, LLMs are used to both decompose tasks into subtasks and to generate task-dependent specifications for each subtask. These specifications are then used by ACL algorithms to guide learning using only sparse reward signals, eliminating the need for dense reward design. We evaluate LEACL on five long-horizon manipulation tasks from the LIBERO benchmark. LEACL achieves better asymptotic performance in terms of the success rates compared to human-designed dense rewards.

Robot policies are typically MLPs mapping observations to actions. Yet robot observations are physical variables, and many action-relevant cues arise not from individual variables but from their interactions; power, inertial effects, contact, slip, and compliance depend on products among observable signals. We introduce PRISM, a policy representation that makes polynomial interactions among observable physical variables explicit, learnable, and compact. Rather than listing all polynomial terms, PRISM uses a factorized polynomial module to expose higher-order interaction features efficiently. In reinforcement learning, it keeps the standard MLP backbone but applies a gradually activated element-wise polynomial function after it. In imitation learning, it replaces linear proprioceptive conditioning in Diffusion Policy with a polynomial layer trained end-to-end. Across humanoid locomotion and contact-rich manipulation, PRISM improves performance over standard MLP policies and larger MLPs with matched capacity, showing that interaction structure cannot be replaced by capacity alone. It also yields sensorless compliant behavior without force, wrench, tactile input, contact labels, or admittance control. These results suggest that polynomial representations should become a standard architectural choice for embodied motor control. The project page is available at https://lsh3163.github.io/prism/

Passive and active photonic components are central to the continued development of astronomical photonic integrated circuits (PICs), although achieving broadband achromatic performance remains a significant challenge. This work presents broadband evanescent tri-couplers, tapered directional couplers, and a chromatically controlled achromatic intensity modulator for astronomical interferometry in the J- and H-bands. Silicon nitride and silicon oxide provide complementary low-loss platforms, while customised tapered components enable broadband operation across the 0.95-1.8 microns range. For the silicon nitride platform developed by STMicroelectronics, tapered tri-couplers and directional couplers were designed as replacements for conventional components in a pairwise beam combiner for the PLANETS project. Optimised tapered tri-couplers achieve less than 1% excess loss across the J-band, while tapered directional couplers provide broadband 40:60 splitting suitable for beam combination. For the silicon oxide platform from Enablence, a two-dimensional tapered tri-coupler was investigated for nulling interferometry. The device provides broadband starlight suppression while limiting exoplanet throughput loss to less than 2.2% across the H-band and simultaneously retaining broadband phase-sensing capability for fringe tracking. The chromatically controlled achromatic intensity modulator is introduced as a combination of a tapered directional coupler with a thermo-optic phase shifter, the device provides programmable broadband intensity control for applications including null-depth balancing in interferometric PICs. Future work will extend these concepts to lithographically fabricated chalcogenide glass platforms operating in the mid-infrared, enabling compact photonic beam combiners for future ground- and space-based nulling interferometers targeting Earth-like exoplanets.

Automated photogrammetric inspection emits metric measurements from a 3D reconstruction whose own correctness is normally unknown without an external survey. Can a reconstruction estimate its own reliability with no ground truth? We formalise a track-leakage-free hold-out protocol: a deterministic subset of images is withheld and each re-localised against only 3D points seen by at least two retained images -- a track-level barrier so a view is never tested against structure it helped create -- aggregated into an mAA confidence. Held-out observations still entered the bundle adjustment that built the trusted structure, so we call it track-leakage-free (a stricter re-mapping variant confirms this at 3/5 deg). Across operational GNSS-referenced captures plus ETH3D, EuRoC and the IMC 2025 benchmark we report four findings. (i) The protocol is computationally well-posed: good reconstructions score millidegree self-consistency. (ii) Thresholded self-consistency saturates and does not track absolute accuracy: confidence stays near 1.00 while true RTK error swings up to 14x within a capture, and the per-capture correlation is sign-unstable across five captures (-0.57 to +0.98; pooled 95% CI spanning zero), so it cannot gate accuracy. (iii) It flags gross failure only when the failure destroys internal consistency: a fragmenting model drops confidence, but a single self-consistent, globally-distorted model evades it -- three of four captures gave one model wrong by 55-106 m at confidence 1.00. (iv) The same dichotomy holds on independent ETH3D and IMC 2025 ground truth. Track-leakage-free hold-out measures internal geometric consistency, not absolute accuracy: neither a substitute for control-point assessment nor a general gross-failure gate. We release the protocol and degradation harness.

Elastic object manipulation (EOM) involves highdimensional, nonlinear, and elastic deformations. The diverse deformation properties of elastic objects substantially expand the relevant state space, requiring extensive exploration to learn accurate manipulation policies for tasks such as slingshot manipulation. While simulation enables large-scale and safe exploration compared to costly and potentially destructive real-world trials (e.g., repeated projectile launches), accurately calibrating elastic behavior between the real world and simulation remains challenging since elastic properties are largely indistinguishable from visual observations alone. To address these challenges, we propose Sling2Sim2Real, a one-shot Real2Sim2Real framework that identifies elastic parameters from a single non-destructive interaction and enables policy learning in simulation. The framework consists of two stages: 1) a multi-start Real2Sim system identification method that exploits parameter covariance to estimate elastic properties, and 2) simulation-based policy learning followed by zero-shot Sim2Real transfer using the calibrated simulator. We evaluate Sling2Sim2Real on a slingshot manipulation task using a Franka Emika Panda arm and elastic bands with diverse physical properties across varying target distances. Experimental results demonstrate that Sling2Sim2Real achieves accurate policy learning and robust generalization while significantly reducing the amount of required real-world interaction.

Scalable reconfigurable photonic integrated circuits require low-loss, high-speed optical control without continuous holding power. Yet widely used thermo-optic tuning and continuously biased electro-optic tuning consume static power and introduce thermal crosstalk or bias drift. Here we demonstrate a monolithic non-volatile photonics platform on lithium tantalate-on-insulator (LTOI). In congruent x-cut lithium tantalate, the switched ferroelectric-domain configuration is retained after the write field is removed. On the same LTOI platform, we demonstrate a waveguide propagation loss of approximately 0.05-0.06 dB/cm and multilevel non-volatile phase tuning in separate devices. The programmed states remain distinguishable through $10^{6}$ write cycles. Weighted segmented electrodes resolve 137 phase positions across a $π$ range, corresponding to an analogue phase-setting resolution of approximately $0.007π$. We further combine non-volatile phase control with high-speed electro-optic modulation to achieve zero-static-power bias control of a >110 GHz modulator and a 59.3 dB extinction ratio after non-volatile trimming. At the system level, an image-edge-detection chip achieves a measured energy efficiency of 3.48 TOPS/W. These results establish LTOI as an integrated photonics platform that combines persistent optical reconfigurability with low-loss routing and high-speed electro-optic modulation.

This paper presents the design and optimization of quadcopter firmware for a low-cost UAV based on the dual-core ESP32-S3 microcontroller, targeting real-time flight control, telemetry, and video transmission within stringent embedded resource constraints. The firmware leverages dedicated task allocation across the two cores: 400Hz control loop on Core 1 implements sensor fusion with a Madgwick filter, a hybrid fixed-point PID controller enhanced by selective derivative scaling and feedforward braking. Core 2 handles 50 Hz telemetry encoding and UDP-based video and data streaming at 25FPS VGA frames, ensuring separation of control and communication workloads to prevent latency interference. Inter-core communication employs zero-copy buffers with mutex protection for minimal overhead and deterministic execution. Custom event-driven threading and RTOS-inspired static scheduling guarantee real-time responsiveness and jitter-free operation despite ESP32-S3's limited memory and floating-point performance. Experimental evaluation confirms that this firmware architecture provides precise, stable flight stabilization and robust telemetry support, enabling resource-constrained ESP32-S3 platforms to handle DIY projects and research with minimal investment and full customization capability.

Digital twins for advanced reactors must invert physical parameters online from noisy, partial sensor streams while the plant is rarely at steady state. Gradient-based inversion is the natural tool for this task, but the forward models in routine use are seldom differentiable end to end, so practitioners fall back on derivative-free filters. We present an end-to-end-differentiable digital twin of a closed-Brayton gas-cooled reactor that propagates reverse-mode automatic differentiation through an implicit differential-algebraic plant model, exposing exact parameter sensitivities. The twin drives an AD-Hessian incremental 4D-Var estimator, benchmarked against ensemble, unscented, and finite-difference variational baselines over a two-by-two matrix that crosses steady with transient excitation and full with partial observation. No single estimator wins everywhere: the unscented filter keeps its best-linear-unbiased advantage on the controlled steady-state corner, while the proposed estimator attains the lowest mean error on the reflector coefficient on the other three corners --- reaching \SI{0.43}{\percent} under transient full observation, roughly an order of magnitude below the unscented filter, and matching the ensemble filter at low-to-moderate noise on the combined transient-partial corner. Every estimator's variance sits within a small factor of the Cramér--Rao bound, so the residual error reflects a deterministic bias floor from the twin's differentiability simplifications rather than statistical inefficiency, and the advantage is one of robustness to the resulting multi-modal loss landscape, which a component ablation localises by regime. Differentiating a first-principles plant model thus makes gradient-based inversion competitive with established filters across a reactor's operating range.

Discovering rare safety-critical failures in autonomous and cyber-physical systems is a fundamental challenge in verification and validation. Existing falsification approaches rely on conditional sampling strategies that factor the joint distribution over environments and system executions, and therefore suffer from multiplicative rarity effects: the simultaneous scarcity of failure-inducing inputs and failure-inducing traces makes exhaustive search prohibitively expensive. This paper develops DiffTilt, a distributional framework that exponentially tilts a diffusion model-induced joint distribution over environments and executions. We show that diffusion-guided sampling admits an exact interpretation as importance sampling in the joint space, where guidance scores induce a KL-optimal reallocation of probability mass towards failure-relevant behaviors. We further show that tilting provably amplifies failure probability and strictly outperforms conditional sampling, which is limited by multiplicative rarity. In this framework, the joint generative model serves as a reusable prior over scenarios and need not faithfully represent the system under test. Expensive system simulations are instead limited to learning a scoring function that characterizes scenario quality, enabling their selective and adaptive use. We study DiffTilt on ARCH-COMP benchmarks, and we propose an additional tractor-trailer benchmark showing the behavior of several approaches when scenario generation is guided by a well-defined specification rather than a reward. The proposed method achieves competitive or improved falsification performance compared to state-of-the-art approaches, with larger gains when specification definition is not limited to STL formulas.

While scaling laws for imitation learning have primarily focused on generalization in open-world settings, the relationship between data and precision in closed-world tasks like robotic assembly remains largely unexplored. This paper systematically investigates this relationship and introduces a novel scaling law. We find that to achieve a fixed success rate, the required number of demonstrations $N$ grows super-exponentially as the target precision $P$ approaches a limit $c$. This relationship is accurately captured by the model $\log(N) \propto 1/(P-c)$. Crucially, we reveal that the limit precision $c$ is not a static physical constant of the task but an emergent property of the entire agent system, including its sensors and expert policy. Through experiments on canonical manipulation tasks, we validate this law and demonstrate that improving system components, such as adding a wrist camera or using a more effective expert, measurably lowers $c$, thus expanding the system's achievable precision. Our work provides a new theoretical framework for precision in robotics and a quantitative metric to evaluate system capabilities. Furthermore, these findings provide a practical methodology for guiding the development and debugging of high-precision manipulation systems.

Batteryless intermittent systems run unattended for years, but power failures erase timekeeping state, corrupting sensing, scheduling, and coordination. State-of-the-art timekeepers infer elapsed time from capacitor discharge; however, the capacitor must be sized for the longest interval measured (so range, energy, and area grow together), and repeated charge-discharge cycling lowers capacitance over time, biasing every estimate further as the deployment ages. We present FLINT, a timekeeper that reads elapsed time from the stochastic retention loss of an array of "broken" Magnetic Tunnel Junctions (MTJs)---spintronic memory cells engineered to lose state predictably. Because the decay timescale is fixed by device geometry, the energy to read it is independent of the interval measured and does not drift with device age. We validate FLINT's array model against 21 fabricated MTJs, then evaluate the full timekeeper in real-device-trace-driven simulation, showing that it tracks over 15 minutes of off-time within 10% error while consuming only 1.03 $μJ$ and occupying under 0.1 $mm^2$---$9.2\times$ the range at $11\times$ lower energy than prior work. It extends to longer intervals at no added cost, and makes $16-52\times$ fewer scheduling errors than an aging capacitor clock over a one-year deployment.

Multi-rendezvous missions for active debris removal and in-orbit servicing incur prohibitive propellant costs when their targets differ in the orientation of their orbital planes, and classical mission design treats the nodal precession induced by Earth's oblateness as a perturbation to be cancelled. This paper presents a drift-aware trajectory optimisation framework that instead exploits that precession as a mission-design resource, benchmarked against the published winning solution to the European Space Agency's Kessler Run trajectory-optimisation competition as a trusted reference. Three methodological contributions are made. First, an enhanced drift orbit shapes the size, shape, and inclination of an intermediate orbit to tune the differential precession rate, substantially cheaper than the altitude-only designs of prior work in the Sun-synchronous regime, where inclination is the dominant lever. Second, a global per-leg time-budget optimiser makes explicit and resolves the coupling between drift-time allocation and downstream nodal geometry that defeats greedy allocation. Third, a sequence-recovery test independently re-derives a removal order consistent with the published reference from the debris catalogue alone. Applied to the full debris campaign on the published mission partition under the competition's transfer-time rules, the framework closely matches the published benchmark cost using only analytical transfer models and minutes of single-core computation. A cross-validation anchored to a widely used flight-dynamics tool quantifies the fidelity limits of the analytic model and bounds the cost of literal execution, and the constructive nature of the algorithm supports rapid in-orbit replanning after a missed manoeuvre.