AI news archive.
10,661 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchDecision Readouts for Text-Mediated Video Anomaly Detection: An Exploratory Evaluation of Jev and Qwen
arXiv:2609.34180v1 Announce Type: new Abstract: How much does the decision readout matter when video-derived textual evidence is held fixed? We evaluate Jev typed decisions and three Qwen readouts on a sparse development…
- arXiv cs.AIresearchEfficient Reasoning via Constrained Optimization in Latent Space
arXiv:2609.34181v2 Announce Type: replace Abstract: Large Reasoning Models (LRMs) have shown remarkable reasoning capabilities, yet they still suffer from overthinking, generating redundant reasoning steps which incur su…
- arXiv cs.AIresearchCASS: Contribution-Aware Structured Sparsity for Model Merging
arXiv:2609.34184v1 Announce Type: new Abstract: Model merging integrates task-specific fine-tuned models into a single multi-task model, but often suffers from parameter interference caused by conflicting task-vector upd…
- arXiv cs.AIresearchPainterBench: A Figural Divergent-Thinking Benchmark for Tool-Using Language Models
arXiv:2609.34195v1 Announce Type: new Abstract: Figural divergent thinking is the ability to develop a given shape fragment into an original drawing. In humans, this ability is assessed with incomplete-drawing tasks. We…
- arXiv cs.AIresearchBehavior-Grounded Semantic Enrichment for Financial Fraud Modeling and Reasoning
arXiv:2609.34211v2 Announce Type: replace Abstract: In financial fraud detection, rich semantic context can provide important evidence for transaction behavior modeling and fraud reasoning. However, public real-world fin…
- arXiv cs.AIresearchEvolving Support Priorities in Empathetic Reinforcement Learning
arXiv:2609.34249v2 Announce Type: replace Abstract: We identify a fundamental mismatch in empathetic reinforcement learning: support priorities evolve with the dialogue state, yet existing methods typically optimize pred…
- arXiv cs.AIresearchQuantaSpike: Short-Window Spike-Driven Quantization for Large Language Models
arXiv:2609.34259v1 Announce Type: new Abstract: Large language models (LLMs) achieve strong performance across many tasks but rely on dense multiply-accumulate (MAC) operations during inference, resulting in high energy…
- arXiv cs.AIresearchQuery Expansion and Key Specialization in Transformer Attention Geometry
arXiv:2609.34273v1 Announce Type: new Abstract: The projection of queries and keys are central to the attention mechanism in Transformer architectures. While they are mathematically symmetric, they play different roles i…
- arXiv cs.AIresearchMoSPR: Histology-to-Gene Expression Prediction with Morpho-Spatial Macrostates and Low-Rank Molecular Programs
arXiv:2609.34280v1 Announce Type: new Abstract: Predicting molecular profiles from histopathology remains challenging because whole-slide images contain spatially organized, heterogeneous tissue patterns, while gene expr…
- arXiv cs.AIresearchDynamical Parameters: An Interpretability Framework for Time-Series Foundation Models
arXiv:2609.34316v1 Announce Type: new Abstract: This work studies a central gap in interpreting time-series foundation models (TSFMs): a dynamical property may be accessible in a hidden state even when the forecast fails…
- arXiv cs.AIresearchTest-Time Scaling via Budgeted Multi-Attribute Verification
arXiv:2609.34322v2 Announce Type: replace Abstract: Verifying LLM-generated answers under a shared computational budget requires jointly deciding which candidates to inspect and which verification attributes to evaluate.…
- arXiv cs.AIresearchKnowing When Thinking Is Not Enough: Teaching Small Reasoning Models to Reason Beyond Their Parametric Knowledge
arXiv:2609.34327v1 Announce Type: new Abstract: Scaling test-time computation is a powerful way to improve language-model reasoning, and is particularly appealing for small reasoning models (sRMs) that are cheap to serve…
- arXiv cs.AIresearchFuzzy Distribution Modeling for Synthetic Tabular Data Generation with Causality Preservation
arXiv:2609.34349v1 Announce Type: new Abstract: Synthetic tabular data generation provides an effective alternative for the training of machine learning models when real-world data is limited or inaccessible. However, th…
- arXiv cs.AIresearchImproving Large Language Models for Code through Runtime Program-State Reasoning
arXiv:2609.34359v1 Announce Type: new Abstract: Large language models receive limited explicit training in reasoning about runtime program states. We study whether training models to reason about runtime program states i…
- arXiv cs.AIresearchCoeF-SFL: Preserving Collaborative Server-Client Learning with Enhanced Communication Efficiency
arXiv:2609.34360v1 Announce Type: new Abstract: Split Federated Learning (SFL) enables resource-constrained clients to participate in collaborative training, but vanilla SFL exchanges smashed data and gradients at every…
- arXiv cs.AIresearchMathematics for and by human cognition: A resource-rational search for bottlenecks in problem-solving
arXiv:2609.34410v1 Announce Type: new Abstract: Human cognitive constraints are generally viewed as limiting factors in problem-solving. We argue that these constraints can instead play a critical role in driving advance…
- arXiv cs.AIresearchOSPD: On-Policy Self-Distillation for Persona-Consistent Dialogue
arXiv:2609.34418v1 Announce Type: new Abstract: Maintaining persona consistency across multi-turn dialogues remains a core challenge for role-playing language models. Off-policy distillation from external teachers incurs…
- arXiv cs.AIresearchCORTEX: Learning to Share and Specialize in Dense Language Models
arXiv:2609.34449v1 Announce Type: new Abstract: Large language models are trained on heterogeneous data mixtures, where different knowledge domains require both shared knowledge and specialization. Existing modular appro…
- arXiv cs.AIresearchWhen Does Structured Knowledge Help Neural Theorem Proving?
arXiv:2609.34460v1 Announce Type: new Abstract: Does structured mathematical knowledge help LLMs prove theorems in Lean 4? If so, for which models, and does the answer vary by problem? Formal libraries such as Mathlib en…
- arXiv cs.AIresearchDoes Model Uncertainty Track Human Ambiguity? Evidence from Multi-Annotator Vision Benchmarks
arXiv:2609.34506v1 Announce Type: new Abstract: Human-model alignment is critical for trustworthy AI-assisted decision-making systems. Yet, most work evaluates model predictions against single ground-truth labels, overlo…
- arXiv cs.AIresearchEOPSA: Efficient On-Policy Self-Distilled Safety Alignment
arXiv:2609.34519v2 Announce Type: replace Abstract: On-Policy Self-Distillation (OPSD) has emerged as a promising paradigm for safety alignment, delivering dense, token-level supervision by distilling from a teacher cond…
- arXiv cs.AIresearchAPOLO: Automatic Prompt Optimization for Ontology Learning
arXiv:2609.34540v2 Announce Type: replace Abstract: Ontology Learning (OL) from text has advanced with the emergence of Large Language Models (LLMs), but it remains challenging due to the limited availability of annotate…
- arXiv cs.AIresearchPersonaManifold: Revealing and Exploiting Curved Geometry in LLM Persona Representations
arXiv:2609.34571v1 Announce Type: new Abstract: Controlling persona in large language models (LLMs) at inference time is important for role-playing, personalized dialogue, and social simulation. Recent methods extract pe…
- arXiv cs.AIresearchNudgeability: Reasoning Models Follow Confidence Signals Without Tracking Their Own Competence
arXiv:2609.34572v1 Announce Type: new Abstract: Reasoning language models that can call tools must decide during inference whether to answer unaided or delegate. Any self-reflection mechanism for this must answer three q…
- arXiv cs.AIresearchDiffusion Subgoal Planning for Long-Horizon Offline Goal-Conditioned Reinforcement Learning
arXiv:2609.34575v2 Announce Type: replace Abstract: Offline goal-conditioned reinforcement learning (GCRL) learns goal-directed policies from reward-free data, but in long-horizon tasks, goal-conditioned value functions…
- arXiv cs.AIresearchCalibrated Uncertainty for Informative Path Planning in Aquatic Environmental Monitoring
arXiv:2609.34577v2 Announce Type: replace Abstract: Informative Path Planning for scalar field reconstruction uses predictive uncertainty to direct sensing vehicles toward maximally informative locations. Gaussian Proces…
- arXiv cs.AIresearchSpeechCritic: Learning a Diagnostic Speech Judge from Limited Human Preferences
arXiv:2609.34582v2 Announce Type: replace Abstract: Human speech conveys rich perceptual information, such as emotion and speaker identity, yet most automatic speech quality judges reduce it to a single naturalness score…
- arXiv cs.AIresearchTULIP: Targeted LLM Unlearning at Layers Identified Per-input
arXiv:2609.34591v2 Announce Type: replace Abstract: Representation-level unlearning modifies the intermediate hidden states of LLMs. Although knowledge is distributed across layers, existing methods operate at a single f…
- arXiv cs.AIresearchA Persistent State for Auditable Mixture-of-Experts Routing
arXiv:2609.34634v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models repeatedly route tokens to sparse subsets of experts, but conventional routers expose no routing-specific record of how cross-layer influenc…
- arXiv cs.AIresearchMechReasoner: A Simulator and Benchmark for Mechanistic Reasoning in Qualitative Physics
arXiv:2609.34636v1 Announce Type: new Abstract: This work introduces MechReasoner, a mechanistic qualitative simulator grounded in confluence-based qualitative physics, together with a benchmark for mechanistic inference…
- arXiv cs.AIresearchOmniTide: Co-Designing Algorithms and Systems for Efficient On-Device Omni-LLM Streaming
arXiv:2609.34653v2 Announce Type: replace Abstract: On-device streaming omni-modal inference safeguards user privacy and eliminates prohibitive per-token API costs, but faces a critical bottleneck: the continuous influx…
- arXiv cs.AIresearchA General Harness for Protein Foundation Model Fitness Prediction
arXiv:2609.34654v1 Announce Type: new Abstract: Accurate fitness prediction is central to protein engineering and understanding sequence-function relationships. With advances in deep learning, protein foundation models (…
- arXiv cs.AIresearchPDE-JEPA: Predictive Representation Learning of Latent Dynamics Modeling for Parametric PDEs
arXiv:2609.34715v1 Announce Type: new Abstract: Physical trajectories contain more than snapshots of a system: they also reveal how its states evolve under governing conditions. However, representation learning for param…
- arXiv cs.AIresearchFrom Human Narrative to Harmonic Structure: A Human-Centered Investigation of Algorithmic Music Generation through the Chord Wheel Diagram
arXiv:2609.34735v1 Announce Type: new Abstract: Contemporary AI-based music generation can produce compositions that satisfy formal requirements of tonality and musical coherence. However, whether musical expression can…
- arXiv cs.AIresearchSeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing
arXiv:2609.34736v1 Announce Type: new Abstract: Large language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query embeddings,…
- arXiv cs.AIresearchPrivacy-Preserving Full-Body Meshing from mmWave Radar via Mesh Foundation Model Supervision
arXiv:2609.34768v3 Announce Type: replace Abstract: Millimeter-wave (mmWave) radar enables privacy-preserving human perception, but the extreme sparsity of point clouds from commercial single-chip sensors (mean ~6.5 poin…
- arXiv cs.AIresearchWhen Do Model Internals Help? Exploring the Role of Representation Engineering in LLM Safety
arXiv:2609.34771v1 Announce Type: new Abstract: Reliable AI safeguards require both control mechanisms that reduce unsafe behavior and monitoring mechanisms that detect safety risks during model interactions. Established…
- arXiv cs.AIresearchBefore the Token Commits: Trajectory-Level Benchmarking of Visual Hallucinations in Diffusion VLMs
arXiv:2609.34772v1 Announce Type: new Abstract: Multimodal diffusion language models generate responses by iteratively unmasking tokens, making each answer the endpoint of a multi-step trajectory rather than an immediate…
- arXiv cs.AIresearchPage-Aware Retrieval-Augmented Generation for EvalLLM 2026: A Five-Variant Study on French PDFs
arXiv:2609.34776v1 Announce Type: new Abstract: We study retrieval-augmented generation (RAG) for questions about French PDF documents when both the answer and its supporting document pages are evaluated. Five system var…
- arXiv cs.AIresearchApplying Language Models in Clinical Medicine: Recent Trends and Perspectives
arXiv:2609.34780v2 Announce Type: replace Abstract: The use and applicability of artificial intelligence (AI) in medical research and clinical practice has received increasing attention in the literature over recent year…
- arXiv cs.AIresearchSTRIDE: Automated Evaluation of Text-to-Trajectory Alignment across Diverse Contexts
arXiv:2609.34799v2 Announce Type: replace Abstract: Language-conditioned trajectory generation is here, but its evaluation has not kept pace. Existing pedestrian trajectory metrics compare trajectories with real-world hu…
- arXiv cs.AIresearchSimulating Respondents, Not Single Questions: Coherent Survey Generation with Large Language Models
arXiv:2609.34828v1 Announce Type: new Abstract: Large language models are increasingly used to simulate response distributions in social surveys. Prior work has achieved accurate population-level simulation for individua…
- arXiv cs.AIresearchBV Loss: Block Verification-Aware Loss for Block Diffusion Speculative Decoding
arXiv:2609.34832v1 Announce Type: new Abstract: Diffusion drafters accelerate speculative decoding by proposing multiple tokens in parallel. Despite recent advances in speculative decoding through sequence-level drafting…
- arXiv cs.AIresearchCan We Trust the Teacher? Decoupled Credit Direction-Magnitude for Self-Distillation
arXiv:2609.34848v1 Announce Type: new Abstract: RLVR provides reliable trajectory-level credit, while OPSD offers dense supervision for token-level credit. This exposes a fundamental coupling when updating step-level cre…
- arXiv cs.AIresearchFrom Soft Targets to Reward Signals: How Assignment and Reward Objectives Interact
arXiv:2609.34850v1 Announce Type: new Abstract: Soft preference targets specify supervision strength, and reward objectives convert that strength into learned reward signals. A central design question remains: how does a…
- arXiv cs.AIresearchAUV-Bench: Aesthetic Understanding and Generation Evaluation for User Interfaces
arXiv:2609.34854v1 Announce Type: new Abstract: Multimodal foundation models are increasingly used for evaluating and generating user interfaces (UIs), often producing seemingly reasonable aesthetic judgments and visuall…
- arXiv cs.AIresearchOn the Limits of Metacognitive Monitoring in LLMs
arXiv:2609.34864v1 Announce Type: new Abstract: Reliable decisions depend on recognizing when an answer may be wrong. In biological cognition, metacognitive monitoring can dissociate from task performance, raising the qu…
- arXiv cs.AIresearchFewer Assumptions by Design: A Reusable Skill for LLM-Assisted Verus Verification
arXiv:2609.34886v1 Announce Type: new Abstract: LLM-assisted Verus verification is a less tedious method to verify Rust implementations, but paired with self-referential structures, e.g., Doubly Linked Lists (DLLs)—not…
- arXiv cs.AIresearchDeShortcut-Align: Decoupling Spurious Shortcuts for Robust Safety Alignment in Large Reasoning Models
arXiv:2609.34896v1 Announce Type: new Abstract: Safety alignment of large reasoning models (LRMs) via supervised fine-tuning (SFT) and reinforcement learning (RL) often yields near-perfect safety scores, yet this apparen…
- arXiv cs.AIresearchDual-Stream Simultaneous Translation via 2D Grid Attention
arXiv:2609.34902v1 Announce Type: new Abstract: Simultaneous machine translation must generate target tokens before the source input is complete. Existing approaches address this through post-hoc read-write policies, lea…