Research

• General in-context learning for embodied agents1 paper raises it · 0 address it

  Enables models to adapt to new tasks and environments without retraining, improving robustness in real-world deployment.

  • Has GPT-6-Astra Achieved Embodied Intelligence?raises
• Cost and latency in embodied control loops1 paper raises it · 0 address it

  High computational and time costs limit real-time applicability of agentic systems in dynamic environments.

  • Has GPT-6-Astra Achieved Embodied Intelligence?raises
• Physical-state verification and failure recovery1 paper raises it · 0 address it

  Agents must reliably detect and correct errors in physical interactions to avoid irreversible failures in real-world tasks.

  • Has GPT-6-Astra Achieved Embodied Intelligence?raises
• Long-horizon task preservation in embodied agents1 paper raises it · 0 address it

  Maintaining progress over extended sequences is critical for complex, multi-step tasks in robotics and navigation.

  • Has GPT-6-Astra Achieved Embodied Intelligence?raises
• Open-source stacks for generalist computer-use agents1 paper raises it · 1 addresses it

  Closed systems hinder reproducibility and systematic study of agent design for real-world computer tasks.

  • OpenGCUA: Toward a Fully Open Stack for Generalist Computer Useraises · addresses

Named by an AI from these papers' abstracts.

• Federated learning50+31.6%
• Agents531+2.3%
• Quantum machine learning240.0%
• Character recognition OCR170.0%
• Sensing communication150.0%
• Frontier LLMs450−7.0%
• Voice AI103−11.2%
• Image segmentation184−11.5%

• arXiv cs.CVOct 19

  Functional Cross-layer Experts Routing for One-step Diffusion Distillation

  Zile Huang, Ser-nam Lim

  One-step distillation compresses diffusion trajectories into a single forward evaluation, yet how internal capacity should be structured within that remaining pass remains largely unexplored. In this setting, identical parameters must simultaneously preserve general generative priors and resolve prompt-specific details without iterative refinement. We formalize this challenge as functional allocation. Our analysis reveals that decoupling representations is beneficial only when expert interference outweighs shared transfer, favoring an always-active shared backbone augmented with conditionally routed residuals over standard isolated MoE architectures. We propose DeMoE, which pairs a persistent shared expert with conditionally routed expert slices across image FFNs, trained under a one-step distribution-matching objective with executed-route supervision. On one-step FLUX.2-4B generation, DeMoE achieves a GenEval score of 0.8220 while activating only 80.5% of the image-FFN width, substantially outperforming dense one-step baselines, matching the 50-step teacher overall, and surpassing it on complex composition and counting. Across COCO-10K, DeMoE establishes state-of-the-art…

  PDF
• arXiv cs.PLOct 19

  Compiling Environment Classifiers to Contextual Modal Types

  Tianyu Chen, Darshal Shetty, Jeremy G. Siek, Chao-Hong Chen, Weixi Ma +2

  Type systems for metaprogramming take one of two approaches to safeguarding the scopes of variables in generated code. They either use Environment Classifiers (EC) to identify scopes by name (Taha and Nielsen, 2003), or they use Contextual Modal Type Theory (CMTT) to match scopes by structure (Nanevski, Pfenning, and Pientka, 2008). In this paper, we connect these two approaches by compiling a calculus with polymorphic environment classifiers, subtyping, and mutable references to a CMTT-based calculus with polymorphic contexts, mutable references, and first-class substitutions. The key idea is to map each environment classifier to a context. We prove that the compilation preserves types and that compilation preserves the behavior of metaprograms. Both proofs are mechanized in Agda. The compilation shows that CMTT is at least as expressive as environment classifiers. We conjecture that the converse compilation is difficult, because environment classifiers impose a stricter regime on the metaprogram.

  PDF
• arXiv cs.LGOct 19

  How Much Does One Wrong Answer Cost? Error Propagation in Constraint Pruning with Noisy LLM Oracles

  Zhengtao Gui, Weining Shen

  Constraint pruning saves queries to a large language model (LLM) or another noisy oracle by skipping each yes/no question implied by answers accepted earlier, for example through transitivity. Entity-resolution systems and LLM-based rankers use this idea. However, their analyses typically assume a correct oracle. We argue that pruning should be judged by its risk when the oracle errs, since an accepted wrong answer spreads to every question it implies. When each answer is wrong independently with small probability $\varepsilon$, the first-order risk is set by how many facts one wrong answer corrupts on average. We consider two kinds of constraints. For implication rules, as in taxonomies, pruning in a random order adds no first-order risk: a wrong answer that would change $k$ facts is accepted with probability $1/k$, so it costs one fact on average. Rules with $s$ premises break this balance: a wrong answer at a premise shared by $r$ rules can cost order $r^{1-1/s}$ facts. For entity resolution with $m$ records of one entity, the first-order risk equals $\varepsilon$ times the Wiener index (the total pairwise distance) of the spanning tree built by the query order. Pivot orders,…

  PDF
• arXiv cs.CLOct 19

  A Calibrated Instrument for Detecting Disposition Drift in Language Models

  Jerry Kaplan

  Before language models are released, they are tested for harmful traits such as willingness to carry out dangerous requests, deception, power seeking, and sycophancy. The safe behavior these tests certify may degrade as models are modified, by topical fine-tuning, new training data, or changes to settings such as the depth of their reasoning. Most work reports such drift as single summary rates scored by another AI model, without the error analysis needed to interpret them. We adapt a calibrated measurement methodology recently developed for assessing how inference optimizations affect the quality of model outputs. Applying it to several misalignment dimensions in two open-weight models, we first establish the instrument's precision, showing that shifts of a few percentage points can be reliably detected. We then induce drift deliberately by fine-tuning on correct and on subtly harmful medical and financial advice; harmful advice induces a far greater shift than correct advice. Most strikingly, a model that reasons before answering can conceal the shift almost entirely in its answers, even though a direct reading of its internal state shows that its disposition has moved. With…

  PDF
• arXiv cs.CVOct 19

  Think Densely, Act Sparsely: Latent Expert Cognitive Chains for Vision-Language-Action Autonomous Driving

  Jie Wang, Guang Li, Zhijian Huang, Jinlong Li, Chenxu Dang +3

  Autonomous driving fundamentally demands dense understanding of visual, semantic, and geometric information, yet existing Vision-Language-Action (VLA) models are typically trained with sparse supervision such as language instructions or trajectory signals. This direct mapping from dense visual observations to sparse representations forces the model to compress or even discard critical scene information essential for safe driving. We argue that this limitation does not stem from insufficient model capacity, but from the lack of explicit mechanisms that encourage the formation of dense world representations in the latent space. In this spirit, we propose a novel paradigm for driving intelligence, Think Densely, Act Sparsely, and instantiate it with an efficient and interpretable framework, LECDrive (Latent Expert Cognitive Chains for Vision-Language-Action autonomous Driving). The core idea is to mimic the hierarchical perception process of humans by embedding a progressive chain of latent expert cognition within the VLA model. Concretely, we inject a compact set of task-specific latent tokens as carriers, enabling the model to internalize dense knowledge from foundational vision…

  PDF
• arXiv cs.AIOct 19

  TimelineAgent: An LLM-based Framework for Query-Centric Event Timeline Extraction Across Multiple Documents

  Qingkai Min, Yunxin Bao, Yifeng Lu, Zhenyu Feng, Yulong Chen +2

  Query-centric cross-document event timeline extraction aims to construct event timelines from multiple documents under diverse query focuses. The task is very challenging due to the heterogeneous event modeling and reasoning requirements arising from both the intrinsic complexity of cross-document timeline construction and the diversity of query focuses. In this paper, we propose TimelineAgent, an LLM-based multi-agent framework featuring a multi-stage reasoning workflow, where heterogeneous event features are organized into reusable skills and orchestrated through a goal-oriented protocol. To facilitate evaluation, we construct a benchmark covering diverse query focuses. Experimental results show that TimelineAgent consistently outperforms existing methods, offering a practical framework for complex real-world timeline construction.

  PDF
• arXiv cs.CLOct 19

  Does Unlearning Survive Collaboration? Robust Unlearning for Shared-Model LLM Teams

  Hengsong Liu, Liangyu Teng, Jingnan Cai, Yuntian Shi, Jingyu Zhang +3

  LLM unlearning is essential for removing sensitive or unwanted knowledge while preserving useful capabilities. As LLM teams are increasingly used to solve complex problems, forgetting must remain effective during collaboration. Despite advances in forgetting effectiveness, efficiency, and robustness to relearning, team deployment remains underexplored. We find that collaboration weakens forgetting: final-answer evaluations miss recoverable evidence in team messages, including complementary clues supporting joint recovery. We establish theoretically and confirm empirically that controlling final answers or individual messages does not ensure low recovery from their combination. To address this challenge, we propose Aggregation-Conditioned Unlearning (ACU). ACU trains a shared model using reconstruction feedback from individual and pooled messages, prioritizes views with higher recovery risk, and uses replacement-based credit assignment while protecting direct and collaborative utility. TOFU evaluates trajectory-level recovery, while WMDP-bio tests native forgetting and retained utility. On TOFU, ACU lowers full-trajectory recovery from 41.8% to 19.3% in ROUGE-L recall and factual…

  PDF
• arXiv cs.LGOct 19

  Neural Video Coding under Limited Delay: Theoretical Optimality and System Design

  Likun Cai, Jun Chen, Ashish Khisti

  Traditional neural video compression (NVC) methods predominantly follow strictly causal predictive-coding paradigm, treating zero-frame delay as a fixed constraint rather than delay as an explicit design parameter. From an information-theoretic perspective, allowing limited frame delay enables a codec to exploit correlations with future frames within the delay window, thereby reducing uncertainty that is inaccessible to strictly causal prediction and potentially improving the rate-distortion tradeoff. In this work, we formalize this intuition through a rigorous rate-distortion analysis builds on the clique factorization of Markov networks, and propose Graph-Factored Sequential Coding (GFSC), a principled framework for sequential video coding under limited delay. To our knowledge, we provide the first theoretical result showing that, for general $K$-order Markov sources with positive distributions, a $K$-frame delay is sufficient to attain joint-coding rate-distortion optimality over a nontrivial distortion region, a guarantee unavailable to zero-delay causal coding in general. In practice, we instantiate GFSC through lightweight neural modules, allowing existing causal predictive…

  PDF
• arXiv cs.CLOct 19

  No Matter Who You Are: Addressivity and Sycophancy in Large Language Models

  Seungjong Sun, Hyunmin Jun, Minsik Yoon, Eungu Lee, Choongwon Kang +2

  Sycophancy, where an LLM becomes overly agreeable, deferential, or flattering toward the user, has been identified as a major failure mode. While post-training has been implicated as a cause, how it internally gives rise to sycophantic behavior remains insufficiently understood. Post-training transforms the model's output from text completion into utterances directed toward an addressee, thereby giving the output \textit{addressivity}. We hypothesize that the accompanying pressure to follow the addressee's instructions and align with their preferences links this addressivity to sycophancy. To test this, we use Sparse Autoencoders to extract addressivity directions from pre-trained and post-trained models and examine their relationship with sycophancy. Steering models with addressivity directions can amplify or suppress sycophancy, and this effect is observed only in post-trained models. Furthermore, comparing the internal mechanisms of both model types confirms that the association between addressivity and sycophancy is strengthened through post-training. Finally, through a controlled post-training experiment, we find that models are sycophantic not toward a specific addressee…

  PDF
• arXiv cs.LGOct 19

  Anatomy of Off-Policy Policy Gradient: Importance Sampling, KL Regularization, and Baselines

  Haoqun Cao, Yurun Yuan, Tengyang Xie

  Reinforcement learning has become a central tool for large language model (LLM) post-training, where policy gradient methods are routinely deployed off-policy, even though vanilla policy gradient assumes on-policy sampling. We study when such off-policy deployment is justified, and what roles its core ingredients, including importance sampling, KL regularization, and baselines, play in handling the resulting distribution shift. We show that for a general class of off-policy policy gradient updates, consistency, in the sense that the optimal policy remains a stationary point, requires one of two structures at the optimal policy: trajectory-level importance weighting, which suffers from the well-known curse of horizon, or KL regularization with a baseline equal to the optimal soft value, instantiated by the grouped mean. Focusing on the KL route, we identify it with the group-relative policy gradient (GRPG) update, establish an equivalence to a trajectory-level Bellman residual minimization objective, and obtain a finite-sample regret bound under all-policy coverage. This also gives a theoretical account of a variant of the empirically successful GRPO objective. We further uncover…

  PDF
• arXiv cs.CLOct 19

  Inside the Scorer: Mechanistic Evidence on Rubric Control in LLM-Based Automated Essay Scoring from 25 Base Models

  Chungche Huang, Po-Ya Angela Wang, Shu-Kai Hsieh

  Automated essay scoring (AES) using large language models (LLMs) offers promising use cases for language assessment, yet high-stakes testing requires scores that are both controllable and verifiable, which current LLM-based scorers lack. Given that, we investigate whether LLM-based scorers form distinct internal representations for individual scoring rubrics and whether these representations can be independently controlled. Probing and steering 25 open-weight base LLMs (0.13B-10.31B parameters; 14 families) on human-scored BESTEP English writing tasks, we find that holistic scores are decodable from model activations for the longer tasks (Parts 2 and 3) and only weakly for the shortest (Part 1). Furthermore, scoring rubrics have partly separable directions in representation space for the longer tasks, though not for short responses (Part 1). However, causal steering reveals a critical limitation: applying a matching intervention (e.g., steering "grammar" on a grammar-degraded response) yields only a negligible advantage (about 2% of the effect) over a mismatched intervention (e.g., steering "vocabulary"), and none that is detectable in any of the 10 testable comparisons across…

  PDF
• arXiv cs.LGOct 19

  Learning to Couple: Context-Adaptive Joint Denoising for Diffusion Language Models

  Zilei Shao, Ian Li, Heng Zhao, Kewei Lian, Guy Van den Broeck +2

  Diffusion language models accelerate generation by predicting multiple tokens in parallel, but factorized denoisers treat simultaneous predictions as conditionally independent. Decoding heuristics reduce the resulting errors by controlling which tokens are committed, but leave the underlying distribution factorized. Joint modeling addresses the source of the problem, yet existing methods attach separately learned layers to frozen denoisers, leaving the backbone unadapted to the joint objective and the coupling parameters fixed across contexts. We propose Codpula, a context-parameterized denoiser combining a Transformer with a tractable probabilistic inference layer. Given a partially masked input, the Transformer produces token potentials that shape per-position predictions and compact gates that modulate the layer's coupling parameters, yielding a context-dependent joint distribution while retaining exact conditional inference. We jointly optimize both components under a conditional denoising objective using a dual-flow multiplicative update that preserves valid probabilistic parameters. At inference, Codpula supports standard heuristics and uses conditional queries to adapt…

  PDF
• arXiv stat.MLOct 19

  Corrected Block Sparse Spectral Optimization: Exact Models, Finite Identification, and Local Dynamics

  Omid Shams Solari

  The block generalized-power method pairs threshold-based support selection with inexpensive Stiefel updates, but the standard block formulation conflates two nonequivalent objectives and leaves inactive components, threshold ties, and post-selection refitting ill-specified. We rebuild the model from the primal. Eliminating each loading column from the normalized loading-ball problem gives a sum of soft-thresholded Euclidean norms (A-SUM); the squared-threshold objective in common use arises from a quadratic lift (A-SQ). We prove both identities, exhibit a two-component instance whose maximizers disagree, and specify explicit conventions for zero-response columns, ties, rank deficiency, and mask-before-refit. For A-SQ we separate the historical full-polar iteration from a safeguarded Riemannian ascent and prove, for the latter, whole-sequence convergence via a sufficient-ascent lemma, a relative-error lemma, and the Kurdyka-Lojasiewicz property, which applies because the soft-threshold objective and the Stiefel constraint are semi-algebraic. A strict threshold margin then forces finite identification of the support and sign pattern. The post-identification map is ordinary power…

  PDF
• arXiv cs.AIOct 19

  Fine-Grained Units, Flat Decisions: Hierarchical Micro-Expert Pruning for Mixture-of-Experts Models

  Zukang Xu, Xinpei Gao, Xing Hu, Zhe Jiang, Dawei Yang

  Micro-expert pruning reduces the capacity of mixture-of-experts (MoE) models by removing coupled channels within routed experts. However, finer pruning granularity alone does not guarantee the preservation of downstream capabilities, as similar language-modeling perplexity can accompany substantially different task performance. We revisit micro-expert mask construction as two distinct decisions: allocating retained capacity across experts and selecting channels within that capacity. We propose UC-RAMP, a two-stage post-training pruning framework that explicitly separates these decisions. CAPE uses cross-domain calibration evidence to redistribute expert widths under an exact layer-wise budget, while RAMP refines channel composition at fixed widths through one-out, one-in exchanges guided by a channel-interaction surrogate. Across three MoE models at 25% and 50% routed-expert channel pruning, UC-RAMP achieves the highest ten-task mean accuracy among evaluated pruning methods in five of six settings. At 50% pruning, it outperforms the strongest evaluated baseline by 8.19, 3.96, and 6.14 percentage points across the three models, respectively. In a single-A100 deployment at 25%…

  PDF
• arXiv cs.CVOct 19

  LEVI: Benchmarking Visual Identity Fidelity with Verifiable Low-Entropy Constructs

  Xiao Ma, Zhonghao Zhan, Hamed Haddadi, Qi Chen

  Modern generative models can produce visually plausible images while violating exact identity constraints. Evaluating such failures requires distinguishing violations of the requested identity from permissible variation in appearance. We introduce Low-Entropy Visual Identity (LEVI), which characterizes visual identities determined by compact facts and explicit rules. The low-entropy requirement concerns identity uncertainty, while appearance may remain diverse. We formulate identity fidelity in terms of the probability that a conditional image distribution assigns to the identities permitted by the prompt. Visual Identity Spaces (VIS) formalize this distinction by specifying admissible identity values and which different realizations express the same identity. Building on this formulation, we develop Verifiable Visual Grammar (VVG), a declarative language whose typed primitives and compositional constraints describe identity-bearing structures. Using this grammar, we build a benchmark for image generation. It assesses each output against the task's identity requirements, drawing evidence from the final pixels. The curated benchmark contains 1,375 tasks spanning 120 base identity…

  PDF
• arXiv cs.AIOct 19

  PFedQFT: Personalized Federated Quantum-Inspired Fine-Tuning for Medical Large Language Models

  Qihan Huang, Lingxia Zhang, Guye Zhang, Zizhu Wang, Xiaowei Mao

  Medical data remain fragmented across institutional silos because of privacy, governance, and ownership constraints. Federated learning enables medical institutions to adapt large language models (LLMs) collaboratively without transferring raw records. However, independently parameterized low-rank adapters incur repeated parameter and communication costs across Transformer projections, while uniform aggregation may limit adaptation to heterogeneous institutional data. We propose PFedQFT, a quantum-inspired framework for personalized federated fine-tuning with a compact shared adapter. Specifically, PFedQFT constructs reusable orthonormal adaptation bases from compact skew-symmetric generators using a truncated matrix-exponential approximation followed by QR orthonormalization. Compatible projections reuse these bases, while a globally shared full core preserves the flexibility of each projection. Each client further maintains a lightweight private residual operator, whose directions are learned globally and whose coefficients remain local. Experiments on seven public medical tasks and multiple LLM backbones show substantial reductions in trainable parameters and per-round adapter…

  PDF
• arXiv cs.LGOct 19

  Evidence Sufficiency and Model Comparisons in MexRegQA: An Item-Level Audit of Mexican-Spanish Regulatory QA

  Fred Hu, Yeqi Wang, Xing Zhou

  Similarity-gated coverage does not establish that retrieved evidence contains the facts needed to answer a question. We audit Mexican-Spanish financial-regulatory QA and find that 35 of 135 covered items lack some or all load-bearing facts. On the post-hoc full-evidence slice (n=98), eleven prompt-matched models score 95-98/98 and none differs significantly from the anchor; this does not establish equivalence. Removing evidence-defective items changes the nominal significance status of 14 of 55 pairwise comparisons. On 56 zero-evidence items, accuracy spans 33.9%-89.3%. Removing evidence from the audited items under an abstention-permitting prompt reduces accuracy by 63-81 percentage points in all ten tested legs, with no recovered items. Building on sufficient-context research, we contribute MexRegQA v1.0: 192 items with provenance and evidence-status labels, frozen retrieval evidence for 191 evaluated items, predictions, judge outputs and offline checks. The audit was outcome-aware, the later model/panel sweep reused the same questions, and all reported scores and labels rely on machine evaluation without independent human validation. We propose releasing evidence-status labels…

  PDF
• arXiv cs.CYOct 19

  Causal Humility: Governing the Authority of Fairness Claims Under Causal Uncertainty

  Khadija Zanna, Rodrigo Ferreira, Frederick L. Oswald, Akane Sano

  AI systems increasingly shape consequential decisions in employment, lending, healthcare, and public services, making fairness an important problem for auditing and governance. Fairness assessment often requires more than identifying disparities in outcomes; it also requires understanding how those disparities arise. Causal approaches to algorithmic fairness address this explanatory problem by representing pathways through which protected attributes and other factors may influence outcomes. Yet causal reasoning in sociotechnical systems remains uncertain: relevant confounders may be unobserved, causal directions contested, and conclusions sensitive to modeling assumptions and representational choices. We introduce causal humility as a governance framework for managing the relationship between epistemic uncertainty and causal authority. Drawing on causal fairness research, epistemology, political philosophy, and science and technology studies, causal humility has four requirements: making consequential assumptions and omissions visible; preserving contestability across defensible causal accounts; recognizing indeterminacy as a legitimate analytical outcome; and maintaining…

  PDF
• arXiv cs.AIOct 19

  Whose Credential Is It? How Owner Attribution in Agent MemoryShapes Cross-User Account Misuse

  Hau-Sing So, Daixian Liu, Carl Hou, Yuhaohang He, Mengzhu Wang +1

  Shared long-term memory lets an LLM agent serving one user inadvertently retrieve another user's credentials, but it is unclear when such exposure leads to actual harm. We introduce a three-stage audit that separates three events: a foreign credential reaches the prompt (exposure), the agent acts with it (behavior), and the other user's account changes (state). In AppWorld, a store that mixes 763 past logins from all users retrieves another user's credential for 84.7 percent of required apps, yet agents rarely act on it. Misuse appears in a stress test where memory lacks the user's own credential and presents a single foreign one as the user's record: seven agents then use it in 80 to 94 percent of runs and change the victim's account in up to 43 percent, sometimes passing the official check. Naming the owner before or inside each credential line prevents most of this: with all else fixed, misuse falls from 92 to 1 of 305 paired runs, also with name-free usernames, though owner-bound notes cost 3 to 5 successful tasks per model on clean memory. A wrongly written owner largely restores misuse; trailing owner tags and added safety instructions help only some models. Agent memory…

  PDF
• arXiv cs.CROct 19

  How Effectively Do Security Skills Guide Coding Agents in Vulnerability Detection in Linux Kernel?

  Haoran Xue, Nikta Akbarpour, Gias Uddin, Song Wang

  Large language model (LLM) coding agents, such as Claude Code, are increasingly used for security auditing, with task-specific skills providing reusable vulnerability-analysis procedures. However, their effectiveness in detecting real vulnerabilities and the reasons for their failures remain unclear. We conduct the first large-scale empirical study of 37 security skills using Claude Code with four underlying LLMs on 100 real Linux kernel vulnerabilities across ten classes, totaling 14800 runs. We evaluate vulnerability detection at the function, statement, and vulnerability-class levels. Results show that 24.4 percent of runs locate the correct function, 18.3 percent identify both the function and responsible statement, and 15.7 percent correctly classify the vulnerability. Our failure analysis reveals limitations in evidence acquisition, state-relation reasoning, code interpretation, candidate verification, investigation coverage, and reporting. Based on these findings, we propose SKILL-SIF, a self-improving framework that iteratively evolves security skills by learning from missed vulnerabilities and failure patterns. We evolve five top-performing skills using the original…

  PDF