Every paper checked — nothing hidden
All papers
Every paper checked for brief #4 of 12–19 Oct 2026 — nothing is hidden. Brief papers first.
All17,313In brief24More (on-topic)3,333Not relevant13,754Not assessed87Older352Could not confirm93Blogs with no paper266All blogs284
| Paper | Topic | Relevance | Feedback |
|---|---|---|---|
| HyrCap: Hybrid Rank-Calibration of Action Proposals for Temporal Event Understanding The paper introduces a rank-calibration framework to improve proposal localization quality for temporal action detection in long untrimmed videos. Ranked #530Preprint0/10 suggested | Multimodal models via arxiv · 15 Oct, 15:06 | On-topic Improves video event understanding and proposal calibration for downstream LLM-driven event analysis. | |
| BANGLAPLAN: A Benchmark for Constraint-Based Planning and Decision-Making in Bangla The paper introduces a 1,000-problem benchmark for constraint-based planning and decision-making in the Bangla language. Ranked #530New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces a benchmark of 1,000 constraint-based planning problems to evaluate LLMs in Bangla. | |
| PhysCodeGym: Benchmarking Coding Agents for Physically Verifiable 4D Reconstruction The paper introduces a benchmark to evaluate coding agents building executable physical simulations from video observations using Blender. Ranked #530New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Examines coding agents using Blender to iteratively reconstruct physical scenes. | |
| Not All Change Is Damage: Update-Matched Unlearning for Large Language Models The paper introduces an update-matched evaluation reference to balance unlearning effectiveness with downstream learning ability in language models. Ranked #531New way to see a problemPreprint6/10 suggested | AI safety and reliability in deployment via arxiv · 17 Oct, 06:18 | On-topic Develops unlearning to remove sensitive data and harmful capabilities while resisting relearning. | |
| AVC-Reward: Event-Centric Audio-Video Consistency Evaluation The authors present AVC-Reward, an event-centric framework to evaluate temporal, narrative, and attribute consistency in generated audio-video content. Ranked #531New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17 | On-topic Proposes a new benchmark and evaluation framework for measuring multimodal audio-video consistency. | |
| Kandinsky S2V 1.0: Audio-Driven Avatar Video Generation The paper presents open speech-to-video models that generate long-form avatar video with expressive motion driven by an audio waveform. Ranked #531Preprint0/10 suggested | Multimodal models via arxiv · 15 Oct, 15:06 | On-topic Introduces multimodal models that generate synchronized video avatars from speech and text. | |
| Interactive Navigation Beyond Reachability: Benchmark and Agentic Framework Introduces a benchmark and agent framework for embodied navigation tasks requiring physical environment interactions like opening doors. Ranked #531New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Develops MultINav-Agent, an agentic framework using VLMs for multi-step interactive spatial navigation. | |
| SPECTRA-VLA: Base-Fine Hyperspectral Token Transfer at Frozen Vision-Language-Action Interfaces Transfers hyperspectral tokens into frozen vision-language-action models via learned continuous residuals without retraining. Ranked #532Preprint0/10 suggested | Multimodal models via arxiv · 15 Oct, 15:06 | On-topic Enriches vision-language-action interfaces with hyperspectral input tokens for multimodal tasks. | |
| Sparse Transformer: a Controlled Testbed for Mechanistic Interpretability The paper introduces a sparse transformer testbed with interpretable neurons to evaluate mechanistic interpretability methods. Ranked #532New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces a controlled sparse transformer testbed to evaluate mechanistic interpretability methods. | |
| PsyMemBehav: Psychology-Informed and Memory-Grounded Two-Level Identity-to-Action Reasoning for Social-Media Behavior Simulation The paper develops a dataset and framework to simulate realistic user social media actions using psychological profiles and retrieved interaction histories. Ranked #532New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Designs memory-grounded multi-step agents to simulate realistic user platform actions. | |
| SkillMarker: A KGW-Based Watermarking Method for Agent Skill Ownership Verification The paper introduces a watermarking method to verify ownership and trace unauthorized copying of agent skill instructions. Ranked #532New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Introduces a watermarking method to verify ownership and protect agent skills from unauthorized leakage. | |
| Why Do VLMs Fail at Remote Physiological Measurement? Rethinking with Physiological Grounding and Numerical Alignment VITAL introduces visual-physiological token alignment to enable vision-language models to perform remote physiological measurement alongside general visual reasoning. Ranked #533Preprint0/10 suggested | Multimodal models via arxiv · 15 Oct, 15:06 | On-topic Integrates physiological signals with vision-language models for multimodal video understanding. | |
| Recognizing Dialogue Ambiguity via Candidate Set Visual Grounding This paper introduces a task and benchmark for tracking candidate sets of referent objects across ambiguous interactive dialogue turns. Ranked #533New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces CandVG, a benchmark evaluating model performance under dialogue ambiguity. | |
| Beyond Policy Rollouts: Embedding-Space Directional Exploration with Anchored Noise for LLM RLVR This paper introduces an embedding perturbation technique to improve exploration diversity during reinforcement learning with verifiable rewards for language models. Ranked #533New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:18 | On-topic Improves policy exploration and reward stability for reinforcement learning with verifiable rewards. | |
| Neos: Learning to Interleave Search and Visual Reasoning Develops a multimodal search framework that learns to turn intermediate visual discoveries into guidance for subsequent search actions. Ranked #533New way to see a problemPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Proposes Neos, an agent learning to interleave tool use, content search, and visual reasoning. | |
| Focus, Align, and Diffuse: Time-Series-Aware Keyframe Diffusion for Cardiac Dynamic Synthesis from Sparse Observations CardioFAD resolves temporal ambiguity in sparse cardiac dynamic synthesis by using ECG-aligned motion-salient keyframe diffusion. Ranked #534Preprint0/10 suggested | Multimodal models via arxiv · 15 Oct, 15:06 | On-topic Aligns ECG time-series signals with sparse visual frames for multimodal synthesis. | |
| KoAGE: Do Large Language Models Know, Learn, and Coin Korean Neologisms? The paper introduces a benchmark to evaluate how well language models recognize, infer, and coin newly emerging Korean terms. Ranked #534New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces KoAGE, a cutoff-aware evaluation benchmark testing neologism acquisition and generation in LLMs. | |
| AI Discovery in Combinatorics: From Problems to Solutions and Back to Bigger Problems Introduces a benchmark of 29 open algebraic combinatorics problems to evaluate algorithm discovery by frontier AI models. Ranked #534New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Provides a scaffold and environment for AI agents searching for combinatorial algorithms. | |
| Bridging Human and Model Attention: Explainability Analysis of CNN, Mamba, and ViT Architectures with Gaze-Based Validation Evaluates explainability methods across CNN, ViT, and Mamba models using eye-tracking gaze data as ground truth. Ranked #534New testPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 12 Oct, 16:28 | On-topic Evaluates explainability and attention alignment with human perception for high-risk AI deployments. | |
| BridgeAfford: Bridging Query-Target Gap for Generalizable 3D Affordance Grounding Introduces a framework that aligns interaction regions across objects to improve 3D affordance grounding on unseen objects. Ranked #535Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Grounds natural language interaction queries onto 3D visual object regions. | |
| JockeyBench: Evaluating LLM Agents on Horse-Racing Decision-Making The paper introduces a horse-racing decision benchmark to evaluate whether agents can extract profitable signals from noisy historical data. Ranked #535New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces JockeyBench, a new benchmark evaluating LLM agents on noisy sequential decision-making. | |
| ADAPT: Audio-Depth Action Planning with Transition-Aware Rewards for Lights-Out Industrial Navigation The paper introduces a navigation framework and benchmark using depth and spatial audio for obstacle avoidance in unlit industrial environments. Ranked #535New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Evaluates embodied audio-depth action planning and closed-loop agent navigation. | |
| From Pixels to Privacy-Sensitive Features: An Information-Theoretic Framework for Reconstruction Bounds Introduces an information-theoretic framework using rate-distortion theory to establish computable lower bounds on feature reconstruction under DP-SGD. Ranked #535New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 12 Oct, 16:29 | On-topic Develops a formal information-theoretic security bound against reconstruction attacks in deployed models. | |
| Full-Duplex Interaction with Speech, Expressions and Gestures Presents a full-duplex conversational model that generates real-time speech alongside synchronized facial expressions and body gestures. Ranked #536Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Presents an end-to-end multimodal model coordinating speech, expressions, and non-verbal gesture generation. | |
| Beyond Equation Search: Conjecture-Guided Discovery with SCOPE and ClosureBench Introduces a multi-agent framework where simulated scientists formulate, test, and revise scientific conjectures alongside mathematical equations. Ranked #536New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Presents ClosureBench, a 60-task benchmark assessing conjecture and closure discovery. | |
| ASIWorld: Benchmarking Autonomous Agents in a Persistent Survival Sandbox This paper introduces a persistent survival sandbox benchmark to evaluate sustained goal execution and program synthesis in autonomous agents. Ranked #536New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Introduces a sandbox benchmark testing autonomous agents' long-horizon planning and execution abilities. | |
| Implementing the Spec Growth Engine: Preventing Spec-Code Divergence, and Growing the Spec with Agents Describes the Spec Growth Engine, which uses deterministic validation and LLM agents to prevent specification-code divergence. Ranked #536New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 12 Oct, 16:29 | On-topic Addresses reliability, human gating, and drift prevention in agentic software development systems. | |
| X-Fly: Learning to Reach and Stop in Full-Chunk Aerial Vision-Language Navigation Proposes a training framework enabling aerial vision-language agents to navigate in full action chunks and stop reliably at destinations. Ranked #537Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Integrates vision-language inputs to control multi-step aerial navigation trajectories. | |
| From Triage to Impact: A Multi-Hop Reasoning Benchmark for Explainable Persian Hate Speech This paper introduces a multi-hop reasoning benchmark to evaluate fine-grained, explainable hate speech analysis in Persian. Ranked #537New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces a multi-hop reasoning benchmark for fine-grained, explainable Persian hate speech analysis. | |
| From Existing Methods to New Research Ideas: Two Perspectives on Recursive Self-Improvement The paper presents a unified taxonomy and design space organizing recursive self-improvement methods for language model agents. Ranked #537New way to see a problemPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 06:17 | On-topic Surveys and formalizes recursive self-improvement mechanisms for language agents updating memory and skills. | |
| Two Clocks: Why Larger Rollout Batches Learn Slower but Fail on Schedule in Reinforcement Learning with Verifiable Rewards The paper develops a framework explaining why large rollout batches in reinforcement learning reduce sample efficiency despite maintaining drift thresholds. Ranked #537Surprising resultPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Analyzes training dynamics and collapse failure modes in RL with verifiable rewards. | |
| Where, How, and When: Query-Guided Grounding for Video Understanding Presents methods for query-guided video grounding across object localization, relationship detection, and temporal moment identification. Ranked #538Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Investigates query-guided video understanding and multimodal grounding across text and video. | |
| Beyond Benchmark Accuracy: A Diagnostic Test for LLM Chemical Prediction The paper proposes a diagnostic test hiding target identity to evaluate structure-property chemical reasoning in language models. Ranked #538New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Presents a diagnostic evaluation framework to benchmark LLM chemical prediction without target identity leakage. | |
| Definition Grounding in Theory-Level Auto-formalization The paper proposes definition grounding to continuously evaluate and revise formal mathematical definitions generated by language model agents. Ranked #538New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 15:06 | On-topic Studies agents that formalize textbook definitions and iteratively refine them for downstream usability. | |
| Known-Voice Cloning and the Limits of Personality Transfer for Character Design This paper investigates whether cloning a familiar voice systematically transfers personality impressions to an embodied virtual character. Ranked #538Surprising resultPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Investigates the risk of social manipulation when deploying virtual characters using cloned familiar voices. | |
| TRACER: A Multimodal Harness for Patient Care in Hospitals Introduces a multimodal harness and benchmark pairing continuous video with clinical records to assist patient monitoring in hospitals. Ranked #539Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Integrates continuous patient video feeds with clinical text for multimodal reasoning. | |
| Counterfactual Gate-Support Decomposition: Separating Selection from Redistribution in Training-Data Filtering The paper proposes an experimental design that separates example selection from mass redistribution to analyze the effects of training-data quality filtering. Ranked #539New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces a counterfactual decomposition framework to rigorously evaluate data filtering. | |
| Neurosymbolic Learning of Interpretable World Models The paper introduces a neurosymbolic world model that learns interpretable concepts from temporal logic constraints for agent planning. Ranked #539New way to see a problemPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 15:06 | On-topic Learns interpretable world models for goal-directed planning agents. | |
| CAPABLE: Capability-Aware Policy Adaptation via Behavioral Latent Encoding Introduces an adaptation framework for frozen vision-language-action policies using residual reinforcement learning and online capability inference during joint faults. Ranked #539New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 12 Oct, 16:29 | On-topic Directly tackles physical reliability and fault recovery in deployed embodied policies. | |
| ReasonUni: Reasoning-Augmented Reconstruction Advances Unified Multimodal Models The paper introduces a reasoning-augmented reconstructive post-training framework that unifies multimodal understanding and image generation tasks. Ranked #540Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Couples text-to-image generation and image-to-text reasoning in unified multimodal models using reinforcement learning. | |
| The Sum of All Summaries: PROPALIGN for Graded Proposition Salience Evaluation The paper proposes a multiple-summary metric and dataset to evaluate graded proposition salience across diverse text genres. Ranked #540New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Introduces PROPALIGN and a benchmark dataset to evaluate proposition salience. | |
| JockeyBench: Evaluating LLM Agents on Horse-Racing Decision-Making The paper introduces a horse-racing decision benchmark to evaluate whether agents can extract profitable signals from noisy historical data. Ranked #540New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 15:06 | On-topic Assesses LLM agent workflows involving priced tools and multi-step decision planning. | |
| Reliability-Aware Future Conditioning for Temporally Robust Robot Manipulation Proposes a reliability-aware conditioning method that mitigates harmful temporal misalignments when using generated video futures for robot manipulation. Ranked #540Surprising resultPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 12 Oct, 16:29 | On-topic Addresses the reliability and robustness of generative model conditioning in deployment. | |
| CogDrift-4D: Hierarchical Multimodal Modeling of Temporal Behavioral Decay for MCI Detection Proposes a hierarchical multimodal transformer to model long-term behavioral changes across multi-session interactions for cognitive impairment detection. Ranked #541Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Introduces a multimodal 4D Transformer combining video and acoustic behavioral features. | |
| Bayesian Learning to Select Fair Classifiers The paper presents an online Thompson Sampling bandit policy to select classifiers that balance predictive accuracy and subgroup fairness under delayed feedback. Ranked #541New way to see a problemPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Presents an online evaluation policy balancing performance and subgroup fairness across agents. | |
| Beyond Equation Search: Conjecture-Guided Discovery with SCOPE and ClosureBench Introduces a multi-agent framework where simulated scientists formulate, test, and revise scientific conjectures alongside mathematical equations. Ranked #541New testPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 15:06 | On-topic Introduces multi-agent scientist frameworks coordinating active experimentation, modeling, and conjecture revision. | |
| Decomposing and Predicting Annotation Disagreement in Harmful Content Across Languages This paper predicts human annotator disagreement on harmful content across multiple languages by extracting interpretable semantic question features with models. Ranked #541New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Examines harmful content moderation, annotator disagreement cues, and moderation reliability across languages. | |
| Automatic Sentence Completion Using Bengali Audio Speech with a Transformer-Based Early Speech Recognition System The paper presents an automated speech recognition and sentence completion system to assist dementia patients speaking Bengali. Ranked #542Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Combines speech audio signals and textual representations for assisted communication. | |
| What Does Disagreement Mean? A Semantic Identifiability Frontier for Pairwise Preference Learning This paper establishes theoretical limits on what forms of annotator disagreement can be identified from pairwise preference feedback data. Ranked #542New way to see a problemPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 15 Oct, 15:06 | On-topic Analyzes how annotator disagreement impacts preference data identifiability, informing evaluation dataset design. | |
| MolHop: Scaffold-Hopping Molecular Design with Agentic Reinforcement Learning The paper presents an agentic reinforcement learning framework that navigates organized chemical spaces for cross-scaffold molecular discovery. Ranked #542New way to see a problemPreprint4/10 suggested | AI agents via arxiv · 15 Oct, 15:06 | On-topic Proposes an agentic framework using reinforcement learning to navigate action spaces for molecular design. | |
| A Sense of Proprioception in Language Models The paper trains language models to report internal states through supervised fine-tuning and on-policy learning. Ranked #542New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Investigates internal state self-monitoring and situational awareness, relevant to model legibility and safety. | |
| Beyond Invariance: Factor-wise Selective Treatment of Shifts for Multimodal Learning Proposes a factor-wise representation framework to selectively treat different distribution shifts in multimodal machine learning. Ranked #543Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:17 | On-topic Develops a representation framework handling distribution shifts across multimodal learning tasks. | |
| Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration The authors present an interactive environment to evaluate how language models integrate exploration, hypothesis generation, and reasoning. Ranked #543New testPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17 | On-topic Introduces a structured interactive benchmark for testing LLM inquiry abilities. | |
| Evaluating Inquiry Learning in LLM-Based Agents: Hypothesizing, Reasoning, and Exploration The authors present an interactive environment to evaluate how language models integrate exploration, hypothesis generation, and reasoning. Ranked #543New testPreprint4/10 suggested | AI agents via arxiv · 17 Oct, 06:17 | On-topic Evaluates LLM agents on multi-step scientific exploration and hypothesis cycles. | |
| Delegating Earth science model calibration to knowledge-guided LLM agents The authors introduce a calibration harness using structured domain knowledge to safely delegate Earth science model parameterization to language model agents. Ranked #543New way to see a problemPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Ensures reproducible and reliable execution when delegating scientific tasks to agents. | |
| Which Tokens to Merge? Diffusion Dynamics for Efficient Image Generation Proposes a training-free token merging method that leverages diffusion dynamics to accelerate image generation while retaining background details. Ranked #544Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:18 | On-topic Investigates dynamics-aware token merging to improve efficiency in multimodal text-to-image diffusion. | |
| Rubric Judgments Lack Uncertainty: Conformal Weighting for LLM Evaluation Proposes conformal calibration to weight criterion-level uncertainty when using rubric-based language model judges for reinforcement learning. Ranked #544New way to see a problemPreprint4/10 suggested | LLM evaluation and benchmarks via arxiv · 17 Oct, 06:17 | On-topic Presents a conformal calibration method to handle uncertainty in LLM-as-a-judge rubric evaluations. | |
| PriorityFEC: Deadline-Aware Streaming FEC for KV-Cache Handoff in Edge Multi-Agent Systems The paper proposes an error-correction method to prioritize and transmit key-value cache tokens across edge-deployed multi-agent systems under latency deadlines. Ranked #544New way to see a problemPreprint4/10 suggested | AI agents via arxiv · 17 Oct, 06:17 | On-topic Optimizes KV-cache handoff for collaborative edge multi-agent systems. | |
| Lost in Contrast: Instruction-Preserving Decoding for Reliable Visual Question Answering This paper introduces an instruction-preserving decoding strategy to prevent visual contrastive decoding from discarding valid abstention responses. Ranked #544Surprising resultPreprint5/10 suggested | AI safety and reliability in deployment via arxiv · 15 Oct, 06:17 | On-topic Addresses reliable abstention and hallucination prevention in visual question answering models. | |
| Not Just a Subject: Capturing Adaptive Scene-Specific Variance for Visual Storytelling Presents a training-free framework that balances identity consistency and narrative scene adaptation for multi-frame visual storytelling. Ranked #545Preprint0/10 suggested | Multimodal models via arxiv · 17 Oct, 06:18 | On-topic Aligns narrative text prompts with consistent image sequences in text-to-image models. |
2,941–3,000 of 17,313← PreviousNext →