AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchBeyond Aggregate Accuracy: Item-Level Transitions in Prompt Optimization
arXiv:2610.19084v1 Announce Type: cross Abstract: Automatic prompt optimization (APO) uses feedback from model responses to improve prompts automatically. However, its effectiveness is usually measured through overall sc…
- arXiv cs.AIresearchDexMimic: A Generalist Dexterous Hand-Object Tracking Policy via Value-Guided Reset Allocation
arXiv:2610.19089v2 Announce Type: replace-cross Abstract: Achieving human-like dexterous manipulation remains a central challenge in robotics. Learning from human hand-object interaction (HOI) datasets offers a promising…
- arXiv cs.AIresearchFrom Copying to Composition: How Language Models Develop In-Context Learning
arXiv:2610.19113v1 Announce Type: cross Abstract: In-context learning (ICL) allows language models to adapt their behavior from instructions and demonstrations in a prompt, but accuracy alone does not reveal how contextu…
- arXiv cs.AIresearchGenesis Steering: Better Activation Steering Through Training Data
arXiv:2610.19114v2 Announce Type: replace-cross Abstract: Activation steering controls large language models at inference time, but its effectiveness depends on the choice of reference data. We introduce genesis steering…
- arXiv cs.AIresearchUnlocking Structured Pruning for Reasoning Language Models via Attention-Preserving MLP Trimming
arXiv:2610.19121v1 Announce Type: cross Abstract: Structured pruning reduces LLM inference costs by removing model components to produce smaller dense models. However, existing methods can cause reasoning performance to…
- arXiv cs.AIresearchPairIT: Autoregressive Transformers for Low-Data Molecule Optimization
arXiv:2610.19123v1 Announce Type: cross Abstract: Generating novel molecules with desired properties is a central challenge in drug discovery. Although foundational work focused on de novo generation, practical use cases…
- arXiv cs.AIresearchImagining Goals with Geometric Horizon Models for Policy Generalization
arXiv:2610.19129v1 Announce Type: cross Abstract: Training from offline data has allowed for substantial progress in domains such as robotics, leading to general-purpose policies that can be easily applied zero-shot or e…
- arXiv cs.AIresearchCAFE: Confining Feature Distillation to Language-Routed Adapters for Compact Multilingual Automatic Speech Recognition
arXiv:2610.19134v1 Announce Type: cross Abstract: Feature distillation is usually expected to add to logit distillation. We find that this expectation does not hold for two common feature losses in compact multilingual a…
- arXiv cs.AIresearchMA-Bench: Measuring Operational Planning in Language Models
arXiv:2610.19138v1 Announce Type: cross Abstract: How well can a language model turn a large, incomplete and partly contradictory body of documents into a plan when no single correct answer exists? We introduce Mission A…
- arXiv cs.AIresearchDiagnosing Low-Confidence Patient Profiles via Localized Harmful Covariate Shift Detection
arXiv:2610.19178v2 Announce Type: replace-cross Abstract: Before deploying clinical prediction models, their reliability must be assessed to ensure safe use. Errors may arise from distribution shifts between training and…
- arXiv cs.AIresearchFinite-sample certification of coverage regret for budgeted document selection under adaptive stratified auditing
arXiv:2610.19207v1 Announce Type: cross Abstract: Selecting a budget-constrained set of documentary fragments, for instance the knowledge base of a retrieval-augmented generation system, raises a statistical question: ho…
- arXiv cs.AIresearchTitan-Llama: Retrofitting Pre-trained LLMs with External Neural Memory Modules
arXiv:2610.19210v1 Announce Type: cross Abstract: Transformers are efficient to pre-train at scale because attention permits parallel computation across sequence positions, but long-context inference requires increasingl…
- arXiv cs.AIresearchGeneralizing Identifiability Results in Causal Discovery from Complete Data to Missing Data Settings
arXiv:2610.19211v1 Announce Type: cross Abstract: While causal discovery methods have matured for complete data, such algorithms for handling missing values remain limited. We establish general sufficient conditions unde…
- arXiv cs.AIresearchDo Syntax Heads Influence Model Behavior?
arXiv:2610.19231v1 Announce Type: cross Abstract: Transformer language models develop attention heads that specialize in syntactic relations, for example, attending from verbs to their subjects. Do these heads play a cau…
- arXiv cs.AIresearchRefined Topology-Pass Scheduling: Damped Trust-Region Compensation for Dynamic k-NN Graphs in SE(3)-Equivariant Structure Prediction
arXiv:2610.19246v3 Announce Type: replace-cross Abstract: Equivariant structure-prediction pipelines alternate coordinate updates with the construction of a k-nearest-neighbour (kNN) graph whose membership is a discrete,…
- arXiv cs.AIresearchCopy-Paste Language Models
arXiv:2610.19247v1 Announce Type: cross Abstract: Transformers are intuitively considered to be efficient at copying from their inputs. We actually show that the ability of LLMs to copy in context can be improved through…
- arXiv cs.AIresearchMIND: Instruction-Conditioned Incremental Grounding for Robots
arXiv:2610.19262v1 Announce Type: cross Abstract: Robots following persistent language instructions must decide which cached room judgments remain valid as their environment changes. We present MIND (Maintenance via INst…
- arXiv cs.AIresearchWhose Data, Whose Benchmark? A User-Centric Framework for Auditing Evaluation Resources
arXiv:2610.19268v1 Announce Type: cross Abstract: Benchmarks shape how model capabilities are measured and interpreted, yet their scores can obscure whose conventions, contexts, and judgments define successful performanc…
- arXiv cs.AIresearchExploring Interpretability in Deep Learning-Based Forest Classification - Separating Similar Forest Classes using Sentinel-2 Time Series
arXiv:2610.19270v1 Announce Type: cross Abstract: The ecological insurance theory posits that diverse forests are more resilient and better able to adapt to changing environmental conditions. Increasing drought and heat…
- arXiv cs.AIresearchDimension-Free Scaling Laws for Size Generalization: An Exponential Separation
arXiv:2610.19279v1 Announce Type: cross Abstract: Modern machine learning models increasingly operate on variable-size structured data, such as sets and graphs. A central requirement is \emph{size generalization}: the ab…
- arXiv cs.AIresearchDTDD: Divergence-Triggered Dynamic Distillation for Reliable On-Policy Supervision
arXiv:2610.19302v1 Announce Type: cross Abstract: On-policy distillation (OPD) provides dense token-level supervision on student rollouts, but this supervision can become unreliable when the student and teacher diverge,…
- arXiv cs.AIresearchLLM-jp-4 Technical Report
arXiv:2610.19331v1 Announce Type: cross Abstract: This paper presents LLM-jp-4, a family of Japanese sovereign large language models (LLMs) developed with an emphasis on transparency, reproducibility, and practical usabi…
- arXiv cs.AIresearchAsymptotic Behavior of Transformers
arXiv:2610.19358v1 Announce Type: cross Abstract: In this paper we study the behavior of transformer networks generating text over long periods of time. We show that even a three-token, single-layer attention-only transf…
- arXiv cs.AIresearchTail Competition Explains Scaling in Best-of-N Sampling for Verifiable Problems
arXiv:2610.19362v1 Announce Type: cross Abstract: Best-of-N sampling is a simple inference-time scaling strategy for large language models, yet its behavior under incomplete reward models remains poorly understood. We st…
- arXiv cs.AIresearchEvaluating Compressed Memory: Dependence, Utility, and Content Benefit
arXiv:2610.19373v1 Announce Type: cross Abstract: Does a language model answer better because its compressed memory contains useful document information? Replacing memory vectors with zeros can lower a trained model's sc…
- arXiv cs.AIresearchELARSpeech: A 30K-Hour Dataset Across 675 Underrepresented Languages for Long-Tail Multilingual Speech Technologies
arXiv:2610.19388v2 Announce Type: replace-cross Abstract: Massively multilingual speech modeling has made remarkable progress in recent years, yet existing training resources and benchmarks still underrepresent much of t…
- arXiv cs.AIresearchKoAGE: Do Large Language Models Know, Learn, and Coin Korean Neologisms?
arXiv:2610.19399v1 Announce Type: cross Abstract: Korean neologisms arise through diverse mechanisms, a demanding test of language models beyond lexical memorization. To enable systematic evaluation, we introduce NeoTrac…
- arXiv cs.AIresearchScaling LLM Inference via Parallelism: Throughput-Latency Trade-off
arXiv:2610.19405v1 Announce Type: cross Abstract: Large Language Model (LLM) serving systems must balance latency and throughput while efficiently utilizing distributed GPU resources. This challenge arises because infere…
- arXiv cs.AIresearchVision-Language Models Integrate Conflicting Evidence Like Biased Bayesian Observers
arXiv:2610.19407v1 Announce Type: cross Abstract: Vision-language models must reconcile conflicting text and images. Prior work has characterized the resulting behavior in terms of modality biases--e.g., a "blind faith i…
- arXiv cs.AIresearchRecursive Formation and First-Class Reuse of Compositional Capabilities
arXiv:2610.19414v1 Announce Type: cross Abstract: We introduce FormCraft, a framework for representing acquired operations as first-class capabilities that can be recursively composed, directly invoked, and selectively r…
- arXiv cs.AIresearchVideoHarness-SRSI: Structured Recursive Self-Improvement for Video Understanding Harness
arXiv:2610.19424v2 Announce Type: replace-cross Abstract: LLM system capabilities depend not only on model parameters but also on the executable harnesses that organize information acquisition, tool use, reasoning, and m…
- arXiv cs.AIresearchHow Do LLMs Understand Legal Philosophy? Diagnosing Legal-Philosophical Reasoning in LLMs
arXiv:2610.19427v1 Announce Type: cross Abstract: Legal reasoning is a fundamental capability for large language models (LLMs) to complete legal AI tasks such as legal judgment prediction. In legal practice, this reasoni…
- arXiv cs.AIresearchMIPS: Mitigating Mode Collapse in Reinforcement Learning for Large Outcome Spaces
arXiv:2610.19457v1 Announce Type: cross Abstract: Expected-reward maximization in reinforcement learning concentrates probability mass on a few outcomes, even in tasks that require many distinct, high-quality solutions,…
- arXiv cs.AIresearchIMU-Bench: Benchmarking Wrist IMU Representations Across Kinematic, Semantic and Clinical Tasks
arXiv:2610.19481v1 Announce Type: cross Abstract: Wrist-worn inertial measurement units (IMUs) provide a common sensing modality for tasks ranging from activity and gesture recognition to pose estimation, inertial odomet…
- arXiv cs.AIresearchRobust Real-Time Monocular Human Pose Estimation with Insole-Based Sensor Fusion
arXiv:2610.19494v1 Announce Type: cross Abstract: Real-time monocular human pose estimation is an important sensing component for human-centered intelligent systems. While recent camera-based methods have achieved impres…
- arXiv cs.AIresearchStructure-Preserving Document-Level Data Augmentation via Discourse Parsing
arXiv:2610.19497v1 Announce Type: cross Abstract: Data augmentation (DA) plays a vital role in improving model performance in low-resource or limited-supervision scenarios within Natural Language Processing (NLP). Existi…
- arXiv cs.AIresearchTyped Decision Models on Tabular Data
arXiv:2610.19536v1 Announce Type: cross Abstract: Typed decision models are language models that answer a question with a fixed set of possible answers by returning a probability for each answer instead of generating tex…
- arXiv cs.AIresearchEvidence-Anchored Overdose Death Prediction from Death Investigation Reports
arXiv:2610.19556v1 Announce Type: cross Abstract: Confirmation of overdose deaths in the United States is often delayed due to the complexities of certifying drug-involved deaths. Death investigation reports represent a…
- arXiv cs.AIresearchTEX-MEX: Similarity-Guided Temporal Contrastive Learning for Time Series
arXiv:2610.19557v1 Announce Type: cross Abstract: Contrastive learning for time series depends on deciding which pairs of inputs should be treated as similar and dissimilar. Because few augmentations reliably preserve me…
- arXiv cs.AIresearchREWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning
arXiv:2610.19561v1 Announce Type: cross Abstract: An early mistake in the generated reasoning can constrain all subsequent reasoning. Repeated-generation methods, like MCMC-based approaches, attempt to improve output qua…
- arXiv cs.AIresearchFrame Evidence:A Benchmark for Reference Frames and Visual Evidence in Multi-Image Spatial Reasoning
arXiv:2610.19564v2 Announce Type: replace-cross Abstract: A spatial answer depends on the reference frame specified in language and its supporting visual evidence. Item accuracy alone obscures whether correctness persist…
- arXiv cs.AIresearchLearning Across Temporal Graph Representation Silos
arXiv:2610.19576v1 Announce Type: cross Abstract: Temporal graph data are recorded in different representations, but whether these differences also determine the scope of learned knowledge reuse remains unclear. This stu…
- arXiv cs.AIresearchBeyond Static Positional Encoding: Bayesian Time-aware Attention for Non-Stationary Long-term Time Series Forecasting
arXiv:2610.19603v1 Announce Type: cross Abstract: Transformers have demonstrated strong performance in long-term time series forecasting (LTSF), yet modeling non-stationary temporal dependencies remains challenging. In t…
- arXiv cs.AIresearchInstruction-Conditioned Motion Control for Robotic Arms: Language-to-Code Decomposition and Reinforcement Learning
arXiv:2610.19617v1 Announce Type: cross Abstract: Instruction-Conditioned Motion Control (ICMC) separates parameter specification, instruction generation, and code synthesis for language-driven robot control, making exec…
- arXiv cs.AIresearchDAGFormer: Global Topology Prediction for Transformer Cross-Layer Connections
arXiv:2610.19625v1 Announce Type: cross Abstract: We propose DAGFormer, a Transformer language model whose cross-layer connections are planned by a jointly trained topology predictor. Beyond the residual connections in a…
- arXiv cs.AIresearchEye, Robot: Learning Expressive Gaze Dynamics via a VLM-Guided Transformer-VAE
arXiv:2610.19635v1 Announce Type: cross Abstract: Gaze coordinates attention in human interaction through coupled eye, head, and eyelid motion. A humanoid robot must choose where to look while generating a suitable movem…
- arXiv cs.AIresearchSparse Autoencoders Are a Shared Coordinate System for Symbolic Matching
arXiv:2610.19636v1 Announce Type: cross Abstract: Large language models (LLMs) hold broad knowledge in one set of weights, but lack interpretability and trustworthy reasoning, especially in high-stakes use cases. A neuro…
- arXiv cs.AIresearchAWARE-RG: Abstraction Aware Explanation of Resiliency Graphs for Cyber Physical Systems
arXiv:2610.19650v1 Announce Type: cross Abstract: Understanding the security posture of cyber physical systems (CPS) requires reasoning across both information technology (IT) and operational technology (OT). However, a…
- arXiv cs.AIresearchHACHIMI-II: From Role-Play to Learner Behavior Modeling for Faithful Student Simulation
arXiv:2610.19654v1 Announce Type: cross Abstract: Simulated students let researchers test tutoring systems and teaching strategies before real learners are available, but only if the simulation reproduces where a particu…
- arXiv cs.AIresearchFunction-Space Temporal Self-Distillation for Policy Optimization
arXiv:2610.19687v1 Announce Type: cross Abstract: Reinforcement Learning with Verifiable Rewards (RLVR) improves reasoning in language models, but scalar outcomes provide coarse supervision over response trajectories. Se…