Artificial Intelligence
Engram Injection: A Study of Temporal Validity in Parametric Language-Model Memory
Sam Siavoshian, Omar Ramadan, Amir K. Saeed et al. (+2 more)
A stored fact may stop being valid when a user changes a preference or an event replaces an earlier one. We study whether a
language model can learn these changes in its parameters and answer correctly from only a question and the current time. Engram Injection stores facts in LoRA adapters and supplies the clock through FiLM conditioning on a frozen Qwen2.5-3B-Instruct model. We compare it with additive clock conditioning, prompting with the full dated history, and retrieving four dated records. We train six matched adapters across three seeds after fixing a sampler that could admit test-age examples into training. The training audit reports no such examples among 127,760 evaluated training inputs. On 62 pairs of questions asked before and after an update, at an elapsed age of seven days excluded from training, both adapter methods answer zero pairs correctly under the primary blinded assistant review with order code 1 held fixed. Full-history prompting answers 59/62 pairs correctly, and
retrieval answers 62/62. Separate experiments show that adapters can store facts and that removable
memory units can support local deletion. Combining fact selection with time-based version selection remains unsuccessful. These results separate learning a fact from learning when to use it.
Read the full abstract ▾
All authors Sam Siavoshian, Omar Ramadan, Amir K. Saeed, Benjamin A. Johnson, Benjamin M. Rodriguez
Artificial Intelligence · Cross-listed from cs.IR
Personalized LLM Recommendation with Semantic IDs: Autoregressive Beam Search vs. Multi-Token Prediction
Sejoon Oh, Fernando Amat Gil, Evan Smothers et al. (+11 more)
Large-scale personalization is shifting from
embedding-table
retrieval to LLM-based recommenders that generate the Semantic ID (SID) of the item to recommend, where each SID is a short sequence of discrete hierarchical content codes. A single instruction-tuned model can then unify
retrieval, content understanding, and
language, and recommend brand-new titles from their content alone using SID; then, how should the model generate the SID? Traditionally, models rely on a single
language modeling head that generates
tokens autoregressively, typically using constrained beam search. In contrast, an emerging body of work explores multi-
token prediction (MTP), where dedicated heads predict all
tokens in parallel, substantially reducing the decoding latency associated with beam search. We report an industry-scale, apples-to-apples comparison of these two generation paradigms: single-head autoregressive decoding vs. MTP. We introduce a new auxiliary ranking loss, defined over the joint SID score, to train the MTP objective. We evaluate the SID training objectives on both quality and training/inference cost, inside a multi-task LLM that jointly does personalized
retrieval, content question-answering (QA), and instruction following over a large streaming-media catalog. Our experiments show single-head beam search consistently beats MTP on most product-critical metrics: ranking and content QA, and trains at roughly 2.6x lower dollar cost on Qwen3-8B scale; MTP wins only on raw inference throughput. Around this finding, we detail an efficient training and inference stack and report end-to-end gains over the previous non-SID baseline equipped with an auxiliary ranking head: +77% item cold-start ranking, +22% content-QA accuracy, +2.8x faster training, with
language metric regression traced to SID vocabulary expansion. We address the
language regression by reinforcement learning post training.
Read the full abstract ▾
All authors Sejoon Oh, Fernando Amat Gil, Evan Smothers, Dan Iter, Mark Thornburg, Dawit Mureja Argaw, Sourabh Medapati, Thea Wang, Ria Verma, Qiuling Xu, Harshit Syal, Ehsan Gholami, Moumita Bhattacharya, Ashish Rastogi
Artificial Intelligence
Attention Sinks in VLMs Are Borrowed, Not Seen
Dengjia Zhang, Alexander Martin, Pengyu Yan et al. (+6 more)
We study how to detect and mitigate visual
attention sinks in
vision-
language models (VLMs). Visual
attention sinks are image
tokens that absorb a large share of
attention over the image regardless of the query or task, and they mislead methods that treat
attention maps as a record of where the model “looked.” Prior work has detected sinks through massive activation dimensions and proposed redistributing their
attention, but none have addressed what makes a
token become a sink. Comparing six candidate detectors on ten VLMs from four families, we find that only the centered cosine similarity between an image
token's key and the key of the
language model's text-sink detects visual sinks. Thus, a visual sink is an image
token whose key aligns with the text-sink's key. Moving a sink's key component along this direction to the median of the image keys sends its
attention back to the text-sink (Are Borrowed) while leaving model answers unchanged~(Not Seen). Building on this, we propose a training-free,
entropy-triggered re-look that first cleans the
attention map, then crops the most-attended region of the cleaned map and inserts it into the
reasoning trace. Cleaning adds negligible computation to the forward pass, and the re-look improves accuracy with a mean gain of $+4.3$ points on fine-detail VQA
benchmarks across model families.
Read the full abstract ▾
All authors Dengjia Zhang, Alexander Martin, Pengyu Yan, Xihan Sun, Jiaqi Li, Benjamin Van Durme, Andrew Yates, Reno Kriz, William Walden
Artificial Intelligence · Cross-listed from cs.LG
LLM Value Functions for Credit Assignment in Agentic RL
Kaijie Zhu, Weixiang Yan, Omar Attia et al. (+3 more)
Reinforcement learning (RL) for large
language model (LLM)
agents remains challenging in long-horizon tasks with sparse outcome rewards, where assigning credit to individual actions is difficult. Methods such as GRPO assign the same trajectory-level advantage to all actions, while critic-based methods require learning an additional value function that can be costly and unstable. We introduce LLM Value Function RL (LVRL), which uses the pretrained knowledge and in-context learning ability of an LLM to estimate policy-aware state values without finetuning a separate critic. LVRL conditions the value model on group context and calibrates its predictions using known values at the initial and terminal states, enabling step-level credit assignment from estimated value differences. We further extend this idea to learning from an expert policy through on-policy
distillation (OPD), introducing LLM Value Function OPD (LVOPD), which focuses teacher supervision on actions associated with decreases in estimated value. We evaluate our approach across diverse agentic
benchmarks: AppWorld, BabyAI, and SWE-bench Verified. Across policy scales (4B to 32B), LVRL consistently and substantially outperforms both GRPO and PPO. On AppWorld and SWE-bench Verified, LVRL yields absolute improvements of up to 13.6 and 8.6 percentage points over GRPO. Notably, on tasks requiring extended exploration like BabyAI, LVRL bridges the gap from total failure to 75.0% success. Furthermore, LVOPD yields consistent improvements over standard
distillation baselines on AppWorld. Finally, Monte-Carlo analysis confirms that our frozen critic accurately identifies semantic pitfalls and recoveries, proving that LVRL extracts robust, policy-aware step-level credit from sparse terminal outcomes.
Read the full abstract ▾
All authors Kaijie Zhu, Weixiang Yan, Omar Attia, Abbas Kazerouni, Alexander Toshev, Andrew Szot
Artificial Intelligence · Cross-listed from cs.CV
Adaptive Look, Flexible Thought: Learning How and How Long to Think in Multimodal Reasoning
Quanlin Chen, Xiuwei Chen, Kun Xiang et al. (+5 more)
The introduction of long chain-of-thought (CoT)
reasoning and external tools has substantially improved the performance of
multimodal large
language models (MLLMs) on general and
reasoning tasks. However, existing fast-slow
reasoning paradigms mainly regulate textual
reasoning length, while visual computation introduced by external tools or visual
tokens is often treated as an auxiliary component. This separation prevents MLLMs from jointly optimizing visual perception and textual
reasoning. To address this issue, we propose L2T (adaptive Look, flexible Thought), a unified
reasoning framework that jointly optimizes the form and depth of computation. L2T first discretizes tool-produced visual states into visual
tokens, bringing visual perception and textual
reasoning into a shared autoregressive space. Based on this unified representation, L2T introduces a modality-aware routing reward to select appropriate
reasoning modes and a difficulty-aware fast-slow strategy to adapt
reasoning depth according to task complexity. Experiments on general perception and
reasoning-intensive tasks demonstrate the effectiveness of L2T. Compared with explicit tool-calling approaches, L2T maintains competitive performance while achieving 1.7–3.6 times higher throughput.
Read the full abstract ▾
All authors Quanlin Chen, Xiuwei Chen, Kun Xiang, Zehua Ma, Chunjin Ye, Boqiang Guo, Xianming Liu, Xiaodan Liang
Artificial Intelligence
Passing Tests, Missing Semantics: Evidence of Interpretive Lock-In in Plan-Based DataFrame Code Generation
Jingjing Liu, Dishu Yang, Hongbo Qin et al. (+1 more)
Planning and revision are intended to bring generated code closer to the task specification. Yet a mistaken interpretation can persist across the plan, implementation, and tests - a pattern we call interpretive lock-in. We investigate this risk on 30 DataFrame tasks derived from TPC
Benchmark DS, comparing Frozen One-Pass, Test-Guided Revision, and PlanAudit-Guided Revision. Test-pass counts rose from 22/30 to 29/30 and 30/30, yet no generated solution matched its stored reference output. Evaluator controls returned the expected outcome in 57/57 cases. Same-engine SQL re-execution reproduced 28/30 references; all 77 test-passing runs on these tasks still failed reference agreement. PlanAudit-Guided Revision improved code scores on 3/30 tasks and test coverage on 11/30, but the median paired change was zero for both measures. Failure analysis revealed tests that omitted requirements, adopted candidate-defined schemas, used nondiscriminating fixtures, or repeated the implementation's incorrect logic. Fresh generation on ten targeted tasks produced 0/30 reference matches and 25/25 false passes among test-passing runs. These findings are diagnostic rather than causal: the pipelines differ in feedback, grounding, call budgets, and editing permissions, and the study uses one model and one primary generation per eligible task. Because pre-adjudication human labels were not retained, inter-rater agreement cannot be reconstructed and the mechanism analyses remain exploratory. Overall, increasingly successful self-generated tests did not establish semantic correctness, even after excluding unreproduced references.
Read the full abstract ▾
All authors Jingjing Liu, Dishu Yang, Hongbo Qin, Hansong Zhang
Artificial Intelligence · Cross-listed from cs.CL
An Explainable System for Offensive Content Detection in Nastaliq Urdu: Corpus Curation, Transformer Benchmarking, and Word-Level Attribution
Muhammad Shahid Iqbal Malik
Automated moderation systems are deployed at scale, yet they remain opaque to the users and moderators affected by their decisions, and their coverage is almost entirely confined to English. For Urdu written in the Nastaliq script, no such system exists, and the annotated resources it would be built on are drawn largely from Twitter/X, where character limits constrain how users write. Interpretability has received little
attention, leaving it unclear what such models actually learn. This study addresses these gaps in three steps. First, we curate and publicly release a corpus of 7604 Nastaliq Urdu Facebook comments, extending an earlier corpus by correcting text-level inconsistencies and re-examining the original labels. Second, we fine-tune four Urdu-capable
transformers under a common grid-search protocol and compare them against twelve feature-classifier baselines on identical data partitions. The best configuration, Multilingual Representations for Indian
Languages (MuRIL), reaches 87.05% accuracy and 87.06% macro F1, improving on the strongest baseline by 5.52 percentage points, a margin that exceeds the 95% confidence interval on test accuracy; the two leading
transformers differ by less than the test set can resolve, and performance is stable across five optimizers. Third, we compute word-level SHAP attributions adapted to right-to-left rendering. Attribution mass concentrates on high-frequency slurs but extends to syntactically neutral function words and to political vocabulary characteristic of the source pages, indicating that aggregate performance may rest partly on lexical and topical regularities rather than a generalizable representation of offensiveness.
Read the full abstract ▾
Artificial Intelligence
Agent Sociodrama: Social-Dynamics-Guided Persona Recasting for LLM Agent Teams
Yirui Qi, Xiaoming Zhang and Jaehong Yoon
Large
language model (LLM)
agents are increasingly deployed as teams, yet they fail in characteristically social ways: information holders stay silent, dominant speakers anchor groups on wrong answers, and norm violations go unsanctioned. We present
Agent Sociodrama, a framework for modeling and validating evolving social dynamics in LLM
agent teams. Under fixed personas,
agents' emotional states and directed, asymmetric relationships are updated every round by a unified external LLM judge and shape their subsequent interactions. Across six LLM backbones and five social scenarios, we assess social fidelity through classical social-theory patterns and agreement with human judgments. The resulting trajectories exhibit structured social patterns and model-specific differences. We then introduce Trajectory-Guided Persona Recasting (TPR), which diagnoses group-dynamic failures from interaction transcripts and social trajectories and recasts
agent personas before rerunning the task. Across four collaborative
benchmarks, TPR improves performance from both heterogeneous and randomly initialized
agent persona compositions, yielding up to 83.3%p of accuracy improvement over the initial casting. On representative distributed-information cooperation and mixed-motive bargaining tasks, we outperform SOTOPIA and AgentVerse in three of four comparisons, by up to 37.5%p in Pareto rate. These results show that validated social trajectories provide actionable signals for repairing LLM
agent teams.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.RO
SurgiCLAW: An Interpretable Atomic-Skill Agent for Precision Robotic Micromanipulation
Angen Ye, junjie xie, Yujia Song et al. (+2 more)
Recent
multimodal large
language models (MLLMs) and
vision–
language–action models have provided
robots with new capabilities for long-horizon task understanding and planning. In surgical robotics, however, end-to-end prediction of low-level actions remains limited by insufficient precision, opaque decisions, and scarce data. This paper presents SurgiCLAW, a hierarchical skill-based framework for precision surgical manipulation. SurgiCLAW confines an MLLM to intraoperative visual understanding, structured explanation, and atomic surgical-skill selection, while delegating physical execution to verifiable deterministic control programs. For
brain–machine-interface implantation, a representative high-precision surgical task, we construct a state-to-skill invocation
dataset and expand it through visual and state perturbations. The model first learns structured skill invocation through supervised fine-tuning and is then optimized using three rewards for action correctness, operation prerequisites, and format validity, improving action selection, operational safety, and output parseability. We evaluate subtask success, offline model comparison, real-
robot performance, and ablations. On a fixed test set, SurgiCLAW achieves 94.6% skill-selection accuracy and 93.0% macro-averaged F1 (Macro-F1). Across 50 closed-loop real-
robot implantation trials, it achieves an 88.0% task success rate.
Read the full abstract ▾
All authors Angen Ye, junjie xie, Yujia Song, Junkai Huang, Dapeng Zhang
Artificial Intelligence · Cross-listed from cs.SE
A Decade in a Day: Synthesizing Agent-Native Interfaces from Human-Oriented Software Evolution
Alan Wu, Wei Tao and Yiwen Guo
Large
language model
agents are emerging as an important new class of users of desktop applications. However, existing desktop interfaces were designed for human end users or application developers, leaving
agents to interpret visual controls or discover and compose low-level software operations. This mismatch can hinder even capable
agents, yet it remains unclear what interface properties make desktop software easier for
agents to use. To study this question, we treat software
evolution as a natural source of interface variation and introduce DeskIF-Bench, a
benchmark of 255 artifact-verified tasks across six desktop applications. By holding the task,
agent, and verifier fixed while varying the software release, we measure how interface changes affect
agent performance. Across two models and six applications, the newest release consistently outperforms the oldest. Trajectory analysis identifies three relevant interface properties: discoverability, granularity, and observability. Guided by these properties, we develop ANIS, a framework that synthesizes
agent-native interfaces from the capabilities of installed applications. On an expanded evaluation of 368 tasks across eight applications, atomic interfaces improve success over native SDK use by 3.8 and 3.5 percentage points with DeepSeek V4 Flash and Claude Haiku 4.5, respectively, while reducing online
token usage by approximately 73%. Goal-level commands further raise success to 74.5% and 67.4%, respectively.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
Hyperbolic Time Series Forecasting with Learnable Decomposition and Manifold Dynamics
Sarang Rajendra Patil, Vishal Fenn and Mengjia Xu
Multiscale structure is central to time series forecasting: temporal signals contain slowly varying trends, intermediate seasonal patterns, and high-frequency residuals. Traditional forecasters model this multiscale structure in Euclidean space, without representing the ordering it induces among components. We identify that cascaded decomposition orders components by their characteristic frequencies, forming a natural frequency hierarchy. This hierarchy aligns with the exponential volume growth of
hyperbolic space, which provides increasing representational capacity with radius and naturally accommodates nested hierarchies. We propose HyperForecast, the first long-term time series forecasting (LTSF) framework operating in
hyperbolic space. HyperForecast embeds decomposed temporal segments on the Poincaré ball where radial distance captures this nested structure, from low-frequency components near the origin to high-frequency components near the boundary. We model geodesic velocity through two
manifold-dynamics variants: (a) autoregressive forecasting via Poincaré residual dynamics, and (b) non-autoregressive forecasting that predicts all future segments in parallel. HyperForecast consistently outperforms existing and state-of-the-art baselines on 9 of 11 long-term forecasting
benchmarks, achieving up to 58% lower mean squared error at long horizons. Further analysis reveals that the learned radial
embeddings recover the frequency hierarchy and that
hyperbolic geometry contributes substantially to forecasting performance.
Read the full abstract ▾
Artificial Intelligence
Google Earth AI Geospatial Reasoning Agent: Capabilities and a Systematic Evaluation
Noam Slonim, Aviv Slobodkin, Gil Fidel et al. (+34 more)
Google Earth AI is a family of geospatial AI models, together with an
agent that orchestrates them. We present this
agent, Google Earth AI's Geospatial
Reasoning agent (GRA), a Gemini-powered research system that combines geospatial data sources, foundation models, and tools to answer questions about our
planet. GRA organizes these into modular domain skills and pairs them with upfront geospatial planning to decompose complex queries into multi-stage workflows. Evaluating such systems is hard: real-world problems admit multiple valid workflows, require data discovery across large catalogs, and rely on evolving ground truth. We introduce two
benchmarks and release their development splits. GeoExpertQueries poses complex analytical problems grounded in practitioner use cases and refined by geospatial experts; each is paired with a consensus golden workflow from three independent experts and scored by a human-validated, tool-agnostic autorater. GeoKTools consists of templated queries with deterministic reference solvers that systematically vary the number of required tool types. We also use two external
benchmarks, CBGB and UnivEARTH; the latter we find to be saturated. To isolate the contribution of GRA's tools and its agentic orchestration, we compare four configurations, each with Gemini 3.1 Pro and 3.8 Flash: a standalone LLM, the LLM with basic tools, the LLM with GRA's skills and tools but without its upfront planning and execution guidance, and GRA itself. With at most basic tools, an advanced model recovers much of the answer on templated queries but satisfies only a small fraction of the expert workflows on open-ended problems; GRA performs well on both. The largest step comes from GRA's skills and tools; on open-ended problems, its planning and execution guidance add a further gain with Gemini 3.8 Flash, letting Flash match or exceed Pro at lower cost.
Read the full abstract ▾
All authors Noam Slonim, Aviv Slobodkin, Gil Fidel, Gal Moshitch, Mimi Sun, Lily Mihalkova, Joydeep Paul, Per Bjornsson, Sydney Lin, Jamie McPike, Samapriya Roy, Utkarsh Patel, Gia Jung, Zachary Luna, Allison Berman, William Cai, Xiao Wu, Mohit Agarwal, Mandar Sharma, Aliza Hoffman, Mark Young, Rama Pasumarthi, Roy Lee, Ruth Alcantara, Luke Barrington, Aaron Bell, Reuven Sayag, Avinatan Hassidim, Yossi Matias, Tyler McHenry, Ronald Ho, Christopher Haire, Oliver Guinan, Niv Efron, Tim Thelin, Gautam Prasad, Shravya Shetty
Artificial Intelligence
Visible but Not Avoided: State-Dependent Re-Proposal of Rejected Candidates in Catalog-Based LLM Optimization Loops
Jianlin Chen, Wenhui Chen and Chi Man Vong
LLM optimization loops often show the proposer the candidates it has already tried and rejected, but leave it to the proposer to avoid them. We ask whether avoidance measured in a static probe predicts avoidance in the states an optimizer's own trajectory produces. In catalog-based loops where three open LLMs improve a GUI
agent by naming one of 40 edits, a re-proposal is identified exactly by the edit's name. Clean base-state probes underpredict re-proposal in the proposers' own loops (macro-averaged loop-minus-probe rate 0.32 [0.27, 0.37]), and proposals of already evaluated harness-edit pairs consume 28-63% of valid proposer iterations. The two settings reject different edits: a probe's rejected edits are mostly ones the proposer did not choose, whereas a loop's are its own earlier choices, which remain preferred when the history is removed. Ordinary history lowers re-proposal for two proposers but raises it for the third, and the effects of failure evidence are model-specific. Stating the rejection status in one sentence lowers re-proposal in every model (-0.16 [-0.22, -0.09] relative to a neutral list of the same IDs). For catalog-based loops with exact edit identities, rejection avoidance should be measured on the rejected sets the loop produces.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.CR
SkillMarker: A KGW-Based Watermarking Method for Agent Skill Ownership Verification
Alan Tang, Lifeng Wang, Gang Wu et al. (+3 more)
Agent skills encode operational procedures and domain expertise as reusable knowledge assets, yet their distribution in plaintext exposes them to unauthorized copying and leakage. However, practical mechanisms for skill ownership verification and leakage tracing remain underdeveloped. Watermark
embedding before distribution offers a potential solution, but structured, compact skill instructions impose stringent requirements for factual and procedural accuracy. To address these challenges, we introduce SkillMarker, a KGW-based watermarking method that jointly targets functionality preservation and strong watermark
embedding. SkillMarker unifies quality-constrained rewriting with signal-oriented candidate search. The former separates essential knowledge from flexible expression to preserve instruction fidelity; the latter explores alternative expressions to obtain a stronger watermark signal under preservation constraints. We release an open-source toolkit that integrates watermark
embedding, quality auditing, and detection to support the protection of
agent skills. SkillMarker targets ownership verification and leakage tracing in distributed skill ecosystems, where content may be copied in whole or in part, or incorporated into other skills. Across four
datasets, SkillMarker negligibly affects downstream task performance on average. In our copy-paste tests, window-based detection consistently identifies copied passages containing as few as 1,536 consecutive
tokens. Even after merging with one or three background skills, SkillMarker retains watermark detection rates of 72.2–77.8%.
Read the full abstract ▾
All authors Alan Tang, Lifeng Wang, Gang Wu, Tao Zhang, Yantao Jia, Xiaofeng Zhu
Artificial Intelligence · Cross-listed from cs.LG
Uncovering the Generative Landscape of Diffusion Models: A Spectral-Guided Quality-Diversity Diffusion Framework
Hannah Janmohamed, Joao Correia, Penousal Machado et al. (+1 more)
Denoising
Diffusion Probabilistic Models have emerged as powerful generative priors across domains ranging from molecular design to high-resolution image synthesis. However, standard
diffusion sampling is inherently biased toward high-likelihood regions, often limiting the diversity of outputs. In many domains, a single, high-likelihood sample is insufficient. Instead, diverse portfolios of high-performing solutions are required, motivating methods that uncover a model's full expressive range. Here we show that integrating Quality-Diversity (QD) optimisation with
diffusion models enables the discovery of diverse, high-performing outputs. We first propose Quality-Diversity
Diffusion, a framework that adapts QD to
diffusion models by directly optimising the model's latent noise and conditioning
embeddings and leveraging
embedding models to quantify diversity. Second, we introduce a novel h-space Spectral Mutation operator, which guides search along directions of maximal semantic sensitivity by analysing principal singular vectors of the model's input-to-activation Jacobian. We evaluate our combined approach, Spectral-Guided Quality-Diversity
Diffusion, on text-to-image generation. Against 12 baselines, our method consistently achieves the lowest FID score, demonstrating an excellent balance of image quality and diversity.
Read the full abstract ▾
All authors Hannah Janmohamed, Joao Correia, Penousal Machado, Antoine Cully
Artificial Intelligence
History Replays Itself: Unleashing Long-Context Capability in Linear Attention
Li Nanbo, Firas Laakom, Yilan Zhang et al. (+10 more)
Linear
attention promises the efficiency needed for long-context
language modelling, yet its fixed-size recurrent state must simultaneously support ongoing computation and preserve information from an ever-growing history. This coupling has reinforced a prevailing view that the state's
memory capacity defines the model's long-context capability. We challenge this presumed ceiling: recurrent computation need not bear the full burden of long-term storage. We introduce NoteBank, a sparse episodic
memory that archives selected text before future queries are known, uses the backbone's native features for writing and addressing, and reads retrieved evidence through its native forward computation via replay. This training-free, plug-and-play interface requires no external
embedding model, separately trained retriever, or separate neural reader. At one million
tokens, it raises Gated DeltaNet's mean accuracy across six RULER
retrieval tasks from 1.7% to 70.3%, with no observed significant
retrieval degradation from 4K to 1M
tokens. This recovery adds $\leqslant$0.025% to analytical FLOPs, with persistent context storage approximately 1/2,600 that of a
Transformer's full MHA KV cache at matched width, depth, and precision. Experiments extend from three 1.3B recurrent backbones to Mamba-2-8B and post-trained Qwen3.5-9B, demonstrating gains on LongBench v1/v2 and MemoryAgentBench. Together, these results show that sparse evidence storage and native replay can substantially expand the long-context capability of pretrained linear
attention models while retaining the efficiency of recurrent computation.
Read the full abstract ▾
All authors Li Nanbo, Firas Laakom, Yilan Zhang, Piotr Piękos, Yimeng Chen, Liangyu Wang, Yifan Xu, Mateusz Ostaszewski, Dandan Guo, Wenyi Wang, Yanning Dai, Yuhui Wang, Jürgen Schmidhuber
Artificial Intelligence
MedHarnessBench: A Unified Benchmark for Healthcare Agent Evaluation
Rui Tang and Hongqiu Wang
Healthcare
agents must transform longitudinal clinical evidence into reliable actions through heterogeneous tools and stateful workflows, conditions under which model-centric systems may lose context, misuse tools, fail to verify consequential transitions, or cross authorization boundaries. Existing medical-
agent benchmarks primarily emphasize terminal task outcomes, making it difficult to determine whether execution failures arise from the
agent model, the runtime harness, or their interaction. We present the Medical Harness Protocol, an outcome-preserving framework that independently evaluates seven healthcare-specific process dimensions: Care Action Fulfillment, Clinical Resource Competence, Patient-Evidence
Alignment, Clinical Workflow Continuity, Clinical Auditability, Closed-Loop Verification, and Clinical Conduct Boundaries. A normalized intervention trace further tracks whether runtime support is available, activated, delivered, adopted by the
agent, and propagated into subsequent behavior. We instantiate the protocol as MedHarnessBench, comprising 342 executable tasks across FHIR/EHR operations, healthcare-administration interfaces, and
multimodal clinical tools. Across more than 16,000 controlled task executions, MedHarness yields broad improvements in terminal task success, intermediate subtask completion, and healthcare-specific process quality. We further evaluate MedHarness against multiple general-purpose harnesses and use controlled single-fault perturbations to verify the sensitivity and discrimination of the seven-dimensional evaluation. These results demonstrate that healthcare-specific runtime support can substantially improve both task completion and execution quality, while revealing failures that terminal outcomes alone cannot capture. Code and resources are available at https://github.com/Tomry1114/Awesome-
Agent-Harness-for-Healthcare.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
What DNA Foundation Models Learn Beyond Sequence Composition
Vincenzo Yuto Civale, Andrew David Bagdanov and Alberto Magi
Do genomic foundation models (FMs) provide predictive information beyond simple sequence composition? We evaluate frozen representations from five DNA FMs against normalized k-mer frequencies across 57 genomic classification
datasets and two exploratory gene-expression regression tasks. Simple k-mer features remain competitive with most models, whereas NTv3 shows a consistent advantage; the ranking is robust to globally fixed and multi-scale k-mer baselines and alternative downstream probes. To test incremental predictive utility directly, we compare global 6-mer features with their concatenation with each FM
embedding under a matched Random Forest protocol. NTv3 improves 46 of 57
datasets (mean $\Delta$MCC = +0.072), while Caduceus provides smaller gains and the other models do not improve mean MCC. We separately decompose
embeddings into a component linearly predictable from 6-mer frequencies and a ridge residual, finding that linear recoverability does not reliably indicate downstream utility. Exact 6-mer-preserving shuffles further reveal task-relevant sequence-order sensitivity, most pronounced for NTv3, although order sensitivity alone does not guarantee predictive improvements. Our results establish compositional features as essential controls and offer complementary diagnostics for evaluating the added value of frozen genomic representations, without making claims about fine-tuning capacity.
Read the full abstract ▾
Artificial Intelligence
Suggested Is Not Superseded: A Pre-Registered Evaluation of a System-One Model as a Memory-Retirement Gate
Xupeng Chen and Chenqian Le
Long-term
memory for assistants needs a gate that decides when a new utterance retires an old stored claim.
Benchmark labels for this step mark a pair as changed or not, so they cannot tell a gate that acts on established change from one that acts on a hint. We write a task contract that separates the two (SUPERSEDED vs. NEEDS-CONFIRMATION vs. COMPATIBLE), label 500 test pairs built from STALE (change pairs plus same-user controls) with three LLM annotators and two human adjudicators (one author and a second annotator; provenance reported), and evaluate Jev, a typed "System One" decision model, under pre-registered hypotheses. Of the 179 test pairs STALE marks as changed, 72 meet the contract's bar for an established change. Jev never retired a claim on 141 same-conversation non-changes. Its failure is premature retirement: the binary gate retires 39.2% of claims whose change is only suggested. A three-way decision cuts this to 15.5% at a 15.3-point detection cost, which misses our pre-registered 10-point budget, and it lands on the binary gate's own threshold frontier; these verdicts hold under three readings of the gold labels. On a held-out vault scored against a second annotator's labels, a cascade that escalated uncertain cases to a
reasoning model and looked best on development data was indistinguishable from simply raising the threshold (premature retirement 26.8% vs. 6.5% on dev). Part of the gap comes from the gold labels: relative to that annotator's reading, LLM-majority gold shares most of its errors with a same-family escalation model, and an identity for binary labels says how much such shared errors hide. The rest is small-sample selection on development data. For builders this means a binary score should not be wired to delete: route suggested changes to pending evidence, and do not trust a cascade selected on development data of this size.
Read the full abstract ▾
Artificial Intelligence
SpecForge: Recursive Self-Improvement for Coding Agents via Behavioral Specifications
Yutong Cheng, Haifeng Chen, Peng Gao et al. (+1 more)
Coding
agents based on frozen
language models can improve their performance by refining the external context they rely on, without updating model weights. We study recursive self-improvement through behavioral specifications: natural-
language code documentation that describes the behavior of program entities and guides
agents in generating correct code. We formulate a self-improvement loop in which an
agent generates code from its current specifications, evaluates it with execution-verified tests, and refines the specifications based on observed failures. The improved specifications then guide subsequent generation, progressively improving the
agent's understanding of the codebase. However, recursive refinement is challenging because program entities are interdependent. Improving the specification of one entity can inadvertently degrade the correctness of others, a phenomenon we call output coupling. This creates a whack-a-mole effect in which local improvements cause regressions elsewhere, preventing consistent progress. We introduce SpecForge, a multi-frontier tree search framework that addresses this challenge with three mechanisms: (i) Pareto-guided state selection preserves complementary specification states and prevents premature commitment; (ii) dependency-constrained entity selection prioritizes callee-before-caller refinement; and (iii) error-diversified expansion explores multiple refinements using different subsets of test failures. On DevEval+, SpecForge achieves the best performance across five backbone models, including a 94.9% solve rate with Claude-4.5-Sonnet and a 5.7-percentage-point average improvement over the strongest baseline. The improved specifications also make the
agent better at tasks outside the loop: they transfer to cross-
language translation and to new-feature implementation, where they outperform the
benchmark-provided requirement documents.
Read the full abstract ▾
All authors Yutong Cheng, Haifeng Chen, Peng Gao, Wei Cheng
Artificial Intelligence · Cross-listed from cs.CL
bharatSetu: a 1M x 22 bridge for indic translation
Haq Nawaz Malik, Nahfid Nissar and Faizan Iqbal
India's Eighth Schedule recognises 22
languages, yet parallel data is uneven across them, and many
language pairs, especially among lower-resource
languages such as Bodo, Dogri, Santali and Kashmiri, lack direct bitext. We present BharatSetu 22-Indic 1M, a row-aligned table of 984,764 English sentences, each paired with machine translations into all 22 scheduled
languages (21,664,808 Indic
cells), generated with the sarvamai/sarvam-translate model family. Because every Indic
cell in a row translates the same English sentence, any two columns form a translation pair: 462 directed Indic-to-Indic pairs, or 506 including English, without direct bitext. We call these pivot-derived pairs and propose a confidence layer that lets users choose how strictly each pair is connected. The corpus is machine-generated and not human-verified; quality is expected to vary by
language, domain and sentence style, and pivot-derived pairs can inherit errors from both sides. We therefore release BharatSetu as a machine layer and propose a human-correction layer keyed to the same row identifiers, so that each corrected column upgrades every pair that uses it, together with an evaluation protocol. We report no translation-quality results.
Read the full abstract ▾
Artificial Intelligence · Cross-listed from cs.LG
RespFocus: Learning Temporal Evidence for Weakly Supervised Respiratory Sound Classification
Yi Tian, Xiaohuan Chen, Yufei tang et al. (+5 more)
Respiratory auscultation provides a low-cost means of assessing pulmonary function, yet transient crackles and sustained wheezes can be difficult to recognize in noisy clinical recordings. Conventional deep learning classifiers often operate as black boxes, offering limited insight into the temporal acoustic evidence supporting their predictions. We present RespFocus, a weakly supervised framework that uses multiple instance learning (MIL) to learn inspectable temporal evidence from cycle-level labels and integrate it into classification. We also introduce PedResp2026, a multicenter pediatric respiratory sound cohort collected in routine clinical environments with adjudicated cycle-level annotations. In classification experiments, RespFocus achieves four-class Scores of 69.55% on public ICBHI 2017 and 72.91% on PedResp2026. Evidence diagnostics and matched temporal interventions support the functional contribution of the learned evidence to model predictions. Together, these findings highlight RespFocus's potential as a tool for automated respiratory sound diagnosis.
Read the full abstract ▾
All authors Yi Tian, Xiaohuan Chen, Yufei tang, Xinyu Lin, Hongliang Li, Chengbin Hou, Degen Huang, Rongshan Yu
Artificial Intelligence · Cross-listed from cs.LG
TabMosaic: Unified In-Context Diffusion for Tabular Generative Tasks
Jayoung Kim, Sanghyun Hong and Noseong Park
Generative modeling for tabular data has advanced to the point where models can synthesize realistic records, recover missing values, and generate samples conditioned on class labels or observed attributes. Despite recent advances, existing approaches either require table-specific training or restrict in-context learning to single-
cell prediction. In this work, we formulate these tabular generative tasks as a single problem: modeling the joint conditional distribution of an arbitrary set of target
cells given the observed
cells of an unseen table. To this end, we turn to
diffusion, whose denoising objective is defined jointly over all target
cells rather than one
cell at a time. Under the pretraining distribution, this objective is minimized by the posterior predictive distribution over the target set. Building on this, we propose TabMosaic, a unified in-context
diffusion model for generating mixed-type target
cells in unseen, potentially incomplete tables without table-specific training. We further show that its pretraining objective bounds the conditional negative log-likelihood and the divergence from the posterior predictive for arbitrary target sets, extending the
Bayesian account of in-context learning from a single-value prediction to joint multi-
cell generation. Across 48 real-world tables, TabMosaic ranks first on all evaluated generative tasks against 20 baseline approaches
Read the full abstract ▾
Artificial Intelligence
FARL: A Fully Automated Pipeline of RL Interface and Solver Design
Tianyu Jin, Liuyu Xiang, Chaowen Yang et al. (+3 more)
Existing reinforcement learning (RL) largely focuses on optimizing learning
algorithms within a pre-defined task formulation, while applying RL to new tasks still relies heavily on human expertise to design the learning pipeline: what an
agent observes, how it is rewarded, which
algorithm it uses, and how training is configured. Automating these choices is challenging because they are tightly coupled. We introduce FARL (Fully Automated Reinforcement Learning), a framework that automates RL design by jointly searching observation and reward programs, learning
algorithms, and training settings. FARL uses training statistics and observed behavior to diagnose failures, propose changes, and test whether they improve learning. Its search preserves diverse candidates and the controls needed for these tests, using a staged budget to screen designs before investing in longer training. Experimental
memory records each finding together with its conditions, so an unsuccessful component can be reconsidered in a different combination. Across expert-withheld and expert-available settings, FARL outperforms the evaluated automated baselines on challenging navigation, locomotion, and manipulation tasks. On Adroit Hand Relocate, FARL achieves 97.1% success at 5M training steps per policy, compared with 0% for Eureka and LiMEN and 1.0% for native Proximal Policy Optimization (PPO) under the same policy training budget.
Read the full abstract ▾
All authors Tianyu Jin, Liuyu Xiang, Chaowen Yang, Xingyang Chang, Pei Pei Li, Zhaofeng He
Artificial Intelligence
Radioactive Feedback: Tracing Models Trained by Label-Only LLM Supervision
Yuejia Li, JunHeng Li, Hao Wu et al. (+2 more)
Conventional
distillation transfers capability through teacher-generated demonstrations. LLM-as-a-Judge knowledge
distillation offers a distinct and increasingly practical alternative. In the label-only variant studied here, the student generates every response and improves from scalar scores alone. Prior work shows that this non-demonstrative feedback can improve student capability, but it creates a fundamentally different provenance problem. No teacher-generated answer enters the student's training set, so any evaluator-specific trace must pass through a low-bandwidth reward signal rather than copied text. Existing watermarks and
distillation detectors do not cover this score-only channel. We introduce Radioactive Feedback, which encodes a registered pattern of behavioral preferences in the rewards themselves. The evaluator applies bounded pointwise adjustments to quality-qualified responses, allowing repeated optimization to transfer the pattern to a student trained by another party. Experiments on code and mathematics show that the registered preferences transfer through scores alone and across
datasets, enabling black-box detection against a fixed reference pool.
Read the full abstract ▾
All authors Yuejia Li, JunHeng Li, Hao Wu, Ke He, Mang Ye
Artificial Intelligence · Cross-listed from cs.CV
Rethinking Human Intent-to-CAD: Parametric CAD Model Generation via Cooperative Multi-Task Alignment and Spatial-Aware Reinforcement Learning
Qingwang Zhang, Jiahao Li and Xiangdong Zhou
Parametric Computer-Aided-Design (CAD) modeling from human intent remains challenging, particularly during the conceptual design stage, where design goals are expressed through incomplete and unstructured modalities (e.g., hand-drawn sketches and textual descriptions). In this work, we rethink the human intent-to-CAD pipeline and propose a unified method that directly maps multi-level human intents to executable codes, without assuming the prior existence of target CAD models. To support our study, we construct HiCAD, the first large-scale
dataset aligning hand-drawn sketches, textual descriptions, and parametric CAD codes. Based on this, we introduce HiCAD, a two-stage framework comprising Cooperative Multi-Task
Alignment to bridge the representational gap between heterogeneous inputs, and Spatial-Aware Reinforcement Learning to enforce
geometric and topological consistency. Extensive experiments demonstrate that our method significantly outperforms existing baselines across multiple tasks, validating its effectiveness and robustness in transforming heterogeneous human intents into high-fidelity parametric CAD models. Our project page: https://zqwlearning.github.io/HiCAD.
Read the full abstract ▾
Artificial Intelligence
Knowledge Interaction in Hypernetwork-based Continual Learning: Conflict and Composition
Jiaqi Zheng, Yi Hu, Yadong Mu et al. (+1 more)
Continual learning aims to adapt pretrained large
language models (LLMs) to evolving data while preserving their pretrained capabilities. A central challenge is catastrophic forgetting: adaptation to new data can impair performance on previously learned knowledge. Since continual learning commonly changes model parameters, the model's predictions may draw on knowledge from the base model, the parameter update, and the input context. Understanding how these sources interact is essential: conflicts reveal whether new knowledge can override outdated knowledge, while multi-hop
reasoning tests whether new knowledge can be composed with existing knowledge to support generalization. We study these interactions in a hypernetwork-based adaptation framework, where a shared network generates adapters from context, allowing a frozen base model to adapt without task-specific updates. We find that, under knowledge conflict, models preferentially follow the context, and their choices depend more on where knowledge is presented than on its content. We also show that context and adapter knowledge perturb the base model's representations along directions orthogonal to its clean hidden states, enabling us to derive steering vectors that control which source the model follows. In multi-hop
reasoning, we find that models fail to reliably compose knowledge across the three sources, with the key breakdown arising from the interaction of the hypernetwork with both the base model and the context. Code and reproduction materials: https://github.com/missswiftie/Knowledge-Interaction-in-CL
Read the full abstract ▾
All authors Jiaqi Zheng, Yi Hu, Yadong Mu, Muhan Zhang
Artificial Intelligence · Cross-listed from cs.MA
Dynamic Conflict Clustering with Lightweight Learning and Strategic Repair for Distributed Constraint Satisfaction
Raziyeh Moghaddas and Fabio Caraffini
Distributed Constraint Satisfaction Problems (DCSPs) model distributed decision-making in multi-
agent systems, where
agents assign values to local variables while satisfying shared constraints. However, many DCSP solvers still rely on broad, repeated peer-to-peer communication and are largely reactive, with limited support for strategic coordination or learning. This increases communication overhead, can lead to stagnation, and reduces the success of hard instances. Hence, we propose Dynamic Conflict Clustering-Incomplete (DCC-Incomplete), a repair-based solver that combines conflict-centred decomposition, structured coordination, strategic repair, and lightweight learning for stagnation recovery. DCC-Incomplete groups the unsolved part of the problem into conflict clusters and restricts communication via a coordination layer to
agents within the same conflict region. It also introduces
memory-guided restart, enabling
agents and clusters to pause, revise, and adapt repairs based on neighbouring conflict states. A thorough evaluation of DCC-Incomplete on distributed graph-colouring
benchmarks with planted satisfying assignments shows that its
memory-guided restart improves the success rate from 37.50% to 61.11%, reduces median runtime from 0.565s to 0.367s, and cuts median communication from 12.493 to 6.503 messages per constraint. These results suggest that dynamic conflict clustering, combined with lightweight strategic and learning-guided repair, improves both communication efficiency and robustness in incomplete DCSP solving.
Read the full abstract ▾
Artificial Intelligence
Recall What Is Missing: Composing Evidence from Long-Term Agent Memory
Qi Feng and Jicong Fan
Under a limited context budget, a long-term
agent must recall not merely
memories that are individually relevant, but evidence that is collectively sufficient for the current query. Conventional
memory access retrieves, scores, and truncates
memories, often spending multiple context slots on redundant facts while leaving complementary evidence behind. We cast the read phase of long-term
memory as query-conditioned evidence composition: each
memory is valued by what it adds beyond the evidence already selected. We instantiate this operation with MemPursuit, a training-free mechanism that uses the residual of a sparse convex query reconstruction as a representation-space proxy for what the current
memory set is missing. By reusing the query and candidate
embeddings already available from
retrieval, MemPursuit progressively composes nested evidence contexts without invoking a
language model or cross-encoder to score each query–
memory pair. The effectiveness of our method is verified on 3 popular
benchmarks in comparison to strong baselines. In fixed-budget LoCoMo experiments, MemPursuit improves end-to-end accuracy by 7.2 percentage points at Top-10 and 5.5 percentage points at Top-20, with consistent gains across multiple
memory frameworks, while achieving a 32.9$\times$ speedup in median reranking latency over fully batched BGE-v2-m3. Our implementation is available at https://github.com/FengQi-HITSZ/MemPursuit.
Read the full abstract ▾
Artificial Intelligence
Big-Big-SycoBench: A Benchmark to Detect Sycophancy Through Model Responses to Valid vs. Invalid Arguments
Marthe Ballon, Arne Vanhoyweghen, Seorin Kim et al. (+11 more)
Sycophancy in large
language models risks amplifying unsupported claims or beliefs across the knowledge ecosystem as these systems increasingly contribute to research synthesis, evaluation, and decision-making. A reliable model should revise its judgement when a user identifies a genuine flaw, but should resist pushback based on fallacies, emotion, or irrelevant appeals to authority. While sycophancy is usually measured on questions with a verifiable answer, the evaluative tasks in which it is most difficult to detect have no such ground truth. We introduce Big-Big-SycoBench, which measures sycophancy through argument discernment: the ability to revise an evaluation after valid criticism while resisting invalid pressure. The
benchmark consists of 450 text artefacts with 5,400 arguments for raising or lowering their quality scores, each reviewed by one or more of twelve human annotators. An LLM scores each artefact twenty times to establish its baseline score and uncertainty. The model is then given a valid or invalid argument and asked to update its score. We only count a revision if it exceeds the model's own baseline variability. We use the resulting score delta to distinguish responsiveness to valid arguments from compliance with invalid ones. Across ten models, responsiveness is 88-95% for seven, 76-77% for Kimi K3 and GLM-5.3, and 32% for Opus 5.5. Compliance ranges from 0.2% for GPT-6 Astra to 95% for GPT-4o. Big-Big-SycoBench is intended to guide the development of models whose judgements can be corrected by sound arguments, but not displaced by unsupported pressure.
Read the full abstract ▾
All authors Marthe Ballon, Arne Vanhoyweghen, Seorin Kim, Francesca Carlon, Lynn Delcon, Melika Mobini, Vincent Holst, Lukas Van de Voorde, Tibo Vanleke, Mathieu Louis, Bert Verbruggen, Brecht F. Verbeken, Andres Algaba, Vincent Ginis