AI news archive.
19,861 articles, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchPrefilling the Reasoning Channel: Output-Prefix Attacks on Reasoning LLMs
arXiv:2609.29775v1 Announce Type: cross Abstract: Large Language Models (LLMs) consume and produce a single sequence of text; hence, if text can be added to the beginning of the LLM's response, i.e., an output prefix, th…
- arXiv cs.AIresearchTimeBraid: Unifying Time Series and Language for Understanding and Forecasting
arXiv:2609.29792v1 Announce Type: cross Abstract: We present TimeBraid, a series of unified time-series and language models that align pretrained language models and pretrained time-series foundation models through inter…
- arXiv cs.AIresearchBenchmarking and Domain Adaptation of Automatic Speech Recognition (ASR) for Adolescent Health Communication in Ghanaian Languages
arXiv:2609.29798v1 Announce Type: cross Abstract: This paper presents an end-to-end study of automatic speech recognition (ASR) for adolescent health communication in three Ghanaian languages (Twi, Dagbani, and Ewe). The…
- arXiv cs.AIresearchSEEK: Skill-Routed Evaluation with Evolvable Knowledge for Industrial Search
arXiv:2609.29803v1 Announce Type: cross Abstract: Search quality evaluation provides essential supervision and diagnostic signals for the development and iteration of industrial search systems. Although large language mo…
- arXiv cs.AIagentsHard Stop: Kernel-Level Preemption and Containment for Rogue Agentic Execution
arXiv:2609.29808v2 Announce Type: replace-cross Abstract: In July 2026, an unconstrained autonomous agent participating in a frontier AI cybersecurity evaluation harness breached its evaluation sandbox, established an ex…
- arXiv cs.AIagentsS2Planner: Multi-Scale Semantic Planner for End-to-End Autonomous Driving
arXiv:2609.29813v1 Announce Type: cross Abstract: We present S2Planner, a trajectory planner that combines three front-facing cameras with ego-motion history and the current driving command. A fine-tuned DINOv3 backbone…
- arXiv cs.AIresearchYour Transformer Can Hold Two Thoughts at Once: Evidence of Linear Superposition in LLMs
arXiv:2609.29845v1 Announce Type: cross Abstract: While Large Language Models (LLMs) rely on highly non-linear components, in this work we demonstrate that they exhibit fundamental linearity: when inputs from distinct te…
- arXiv cs.AIresearchTemplate Ageing and Longitudinal Verification in Fixed-Text Keystroke Dynamics: A Subject-Disjoint Study Across Eight Weeks
arXiv:2609.29851v1 Announce Type: cross Abstract: Behavioural biometric templates are widely believed to degrade as the gap between enrolment and verification grows, but few studies measure this template ageing effect di…
- arXiv cs.AIresearchMulti-Task Learning by using Contextualized Word Representations for Syntactic Parsing of a Morphologically Rich Language
arXiv:2609.29855v1 Announce Type: cross Abstract: We address the challenge of syntactic parsing for Urdu, a morphologically rich language, and present state-of-the-art results for both constituency and dependency parsing…
- arXiv cs.AIagentsWorking with Agentic `Teammates': When a New Organizational Actor Collides with the Human Ecosystem of Work
arXiv:2609.29901v1 Announce Type: cross Abstract: Enterprise AI is transitioning from single-user, reactive tools toward proactive, multi-user 'teammates,' but our empirical understanding of this transition is limited. I…
- arXiv cs.AIresearchMorphIK: Morphology-Conditioned Neural Inverse Kinematics for Unknown Robots
arXiv:2609.29908v1 Announce Type: cross Abstract: Neural models can learn to generate various solutions to the inverse kinematics problem from data, but are usually limited to a single robot. We present MorphIK, a flow-m…
- arXiv cs.AIresearchStructured Pose-Conditioned Flow Matching for Generative 5G CSI Augmentation
arXiv:2609.29912v1 Announce Type: cross Abstract: With the growing demand for privacy-preserving and occlusion-resilient human pose recognition (HPR), 5G channel state information (CSI) offers a promising contactless sen…
- arXiv cs.AIresearchCultural Divergence Preservation: Diagnosing Flattening and Caricature in LLM-Simulated Survey Populations
arXiv:2609.29928v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used as synthetic survey respondents to estimate population response distributions. In cross-cultural survey simulation, eva…
- arXiv cs.AIresearchAn Empirical Study of VLM Pipelines for Long-Document QA
arXiv:2609.29933v1 Announce Type: cross Abstract: Vision-Language Models (VLMs) are increasingly used for long-document processing, where the inputs combine text with charts, tables, figures, and complex layouts. Deployi…
- arXiv cs.AIresearchWhen Temporal Perturbations Act Like Sensor Biases: Label-Free Auditing of Wearable Activity Recognizers
arXiv:2609.29937v1 Announce Type: cross Abstract: Wearable human-activity recognition (HAR) models operate across sensors, subjects, and backbones, yet a smooth waveform may appear temporal while exploiting a persistent…
- arXiv cs.AIresearchMind What Matters for Reasoning: Aligning Cross-Modal Attention via Selective Probability Mass Concentration
arXiv:2609.29940v1 Announce Type: cross Abstract: Multimodal large language models (MLLMs) achieve strong performance on visual reasoning tasks, yet remain prone to hallucinations and over-reliance on language priors, of…
- arXiv cs.AIresearchTracking States or Tracking Cosets? An Algebraic Account of Learned State Tracking
arXiv:2609.29951v1 Announce Type: cross Abstract: State tracking requires composing a sequence of updates, but accuracy alone does not reveal what a model has learned. We study neural networks trained to predict the runn…
- arXiv cs.AIresearchBeyond Average Safety: Chance-Constrained LLM Fine-tuning
arXiv:2609.29960v1 Announce Type: cross Abstract: Fine-tuning large language models on new objectives can improve helpfulness, instruction following, or domain-specific performance, but it can also induce regressions on…
- arXiv cs.AIresearchADATEX4D: adaptive texture capacity allocation for 4D gaussian splatting
arXiv:2609.29963v2 Announce Type: replace-cross Abstract: Textured Gaussians improve local appearance capacity, but assigning the same texture resolution to every primitive wastes storage on low-detail or weakly visible…
- arXiv cs.AIagentsWorld Action Agent: Harnessing VLMs for Robot Manipulation via World Action Rehearsal
arXiv:2609.29964v1 Announce Type: cross Abstract: General-purpose vision-language models (VLMs) bring broad knowledge and spatial reasoning to robot manipulation, yet existing systems either use them indirectly, to predi…
- arXiv cs.AIresearchLearning Better Reasoning for Generative Recommendation with Semantic IDs
arXiv:2609.29973v1 Announce Type: cross Abstract: Generative recommendation reformulates item retrieval as sequence generation, allowing a unified model to directly generate the next item from a user's interaction histor…
- arXiv cs.AIresearchFrom Interests to Semantic IDs: Retrieval-Grounded Credit Assignment for Generative Recommendation
arXiv:2609.29983v1 Announce Type: cross Abstract: Semantic IDs (SIDs) encode each catalog item as a short token sequence, enabling generative recommenders to predict the next item autoregressively. Reasoning-enhanced var…
- arXiv cs.AIagentsGuardrails or Roadblocks? Effects of Pedagogical Style and Context Awareness in AI Teaching Assistants for Programming
arXiv:2609.29995v1 Announce Type: cross Abstract: AI teaching assistants (AI TAs) backed by large language models (LLMs) and pedagogical guardrails are increasingly being integrated into programming courses, providing st…
- arXiv cs.AIresearchGHOST-Q: Towards Studying Grounding Hallucinations Overlooked Under Same-score TradeOffs in Quantized VLMS
arXiv:2609.29999v1 Announce Type: cross Abstract: Post-training quantization of vision--language models (VLMs) is typically assessed through aggregate task accuracy and memory savings, but preserving a headline score doe…
- arXiv cs.AIresearchAutomated Regulatory Compliance Question Answering in Financial Services with Domain-Adapted Retrieval-Augmented Generation
arXiv:2609.30009v1 Announce Type: cross Abstract: Financial institutions operate under dense, frequently amended rulebooks, and answering a compliance question correctly requires not only fluency but verifiable grounding…
- arXiv cs.AIresearchLow-Cost Assays for Measuring Model Behavior Across Vendors and Releases
arXiv:2609.30012v1 Announce Type: cross Abstract: Language models advise people, keep them company, and write software while they sleep. Measuring what they do is hard: behavior has to be sampled repeatedly across models…
- arXiv cs.AIresearchCanopy: Exploiting Piecewise Smooth Tree Priors for Multi-Fidelity Bandits
arXiv:2609.30017v1 Announce Type: cross Abstract: Many LLM inference problems, including model routing, prefix-cache management, prompt trimming, and test-time search, can be viewed as optimization over a tree. This stru…
- arXiv cs.AIagentsEra by Eon: Benchmarking Enterprise Agents on Hidden Knowledge
arXiv:2609.30055v1 Announce Type: cross Abstract: In the Era by Eon benchmark, each question states the rules for its answer, and code computes the answer from a generated company's data. When agents can run code, the fo…
- arXiv cs.AIresearchCan Labor Markets Function in the Age of AI? The Evaluation Bottleneck in Hiring
arXiv:2609.30058v1 Announce Type: cross Abstract: AI-assisted job-search tools have become increasingly popular by making it easier to find and apply to jobs. But by making it easier for applicants to generate and tailor…
- arXiv cs.AIagentsKernelOPT: Dispatch-Aware Agentic Search for GPU Kernel Optimization
arXiv:2609.30059v2 Announce Type: replace-cross Abstract: Deep learning inference and training performance depends critically on GPU kernel efficiency. Modern compilers such as PyTorch Inductor automatically generate GPU…
- arXiv cs.AIresearchHow Reproducible Are Evaluation Conclusions? A Self-Audit of LLM-Inferred Prompt Structure
arXiv:2609.30074v1 Announce Type: cross Abstract: Evaluations of LLM systems routinely average over small prompt sets and report models as a ranked table. We ask how much confidence such a table deserves, using LLM-based…
- arXiv cs.AIresearchReachability-Based Formal Verification of Graph Neural Networks with Node and Edge Features
arXiv:2609.30079v1 Announce Type: cross Abstract: Graph neural networks (GNNs) have become a prominent approach for developing fast, topology-aware surrogates in electric power systems, supporting tasks such as power flo…
- arXiv cs.AIresearchAT-SKM-Net: An Accelerated Trainable Sampling Kaczmarz-Motzkin Framework for Linear Hard-Constraint Feasibility on Dynamic Graphs
arXiv:2609.30088v1 Announce Type: cross Abstract: Graph-structured optimization with linear constraints is fundamental to critical infrastructure but faces scalability limits due to massive strict hard constraints and hi…
- arXiv cs.AIresearchAccelerating Video Diffusion via Training-Free Trajectory Routing
arXiv:2609.30096v1 Announce Type: cross Abstract: Video diffusion is computationally expensive, as it requires executing a large model across many denoising steps. Even with step-distillation, inference remains expensive…
- arXiv cs.AIresearchR-DEIM Net: An Efficient Rationale-Augmented Dual-Expert Interaction Model for Paraphrase Detection
arXiv:2609.30100v2 Announce Type: replace-cross Abstract: Recent advances in paraphrase detection reveal a fundamental trade-off: large language models achieve high accuracy but require high computation, while efficient…
- arXiv cs.AIresearchDoes a model's stated reason for rejecting a candidate do any work?
arXiv:2609.30151v1 Announce Type: cross Abstract: Asked to choose between candidates and explain the choice, a language model often rejects a rival by naming a fact its profile lacks: no director, no date of death. That…
- arXiv cs.AIresearchUnderwater C3-JEPA: An Object-Centric Cross-View World Model for ROV Salvage
arXiv:2609.30214v1 Announce Type: cross Abstract: We present Underwater C$^{3}$-JEPA (cross-view, control-conditioned, context-extended), an object-centric multi-view predictive world model for near-field heavy-load unde…
- arXiv cs.AIagentsInstrumental Monitor Evasion Emerges Under Ordinary Task Pressure
arXiv:2609.30217v1 Announce Type: cross Abstract: A central concern in AI safety is that agents may treat oversight as an obstacle when it conflicts with completing their goals. We study instrumental evasion, the propens…
- arXiv cs.AIresearchMinimally Invasive Steering of Language Models
arXiv:2609.30218v1 Announce Type: cross Abstract: Pre-logit steering adapts a frozen language model to a test-time reward by adding vectors to its final hidden states. Unregularized reward optimization can substantially…
- arXiv cs.AIresearchRequirement-Bound Verified Commissioning: A Frozen Four-Billion-Parameter Local Model as a Candidate Generator under an External Acceptance Layer with Verification and Release Authority
arXiv:2609.30219v1 Announce Type: cross Abstract: An acceptance protocol is developed for sensor-coordinate and polarity binding in mechatronic commissioning. Candidate generation is separated from release authority. Req…
- arXiv cs.AIresearchTrackEverything: Long Horizon Dense Tracking via De-Duplicating 3D Scene Representations
arXiv:2609.30222v1 Announce Type: cross Abstract: Existing point tracking models face a fundamental tradeoff: they can either track a sparse set of query points over long horizons, or track all points across only short c…
- arXiv cs.AIresearchPoEM: Predicting RL Outcomes from Existing Policies
arXiv:2609.30226v1 Announce Type: cross Abstract: Foundation models are post-trained with reinforcement learning (RL) to maximize specific rewards, such as human alignment, correctness, or instruction following. This pos…
- arXiv cs.AIresearchTo Trust or Not to Trust: Retrieval-Augmented Fact Checking in Speech
arXiv:2609.30227v1 Announce Type: cross Abstract: Online misinformation increasingly appears in spoken formats such as news clips, podcasts, interviews, political speeches, and social media videos, creating a need for fa…
- arXiv cs.AIagentsCoding Agents for Generalized Task and Motion Planning Problems
arXiv:2609.30233v1 Announce Type: cross Abstract: Task and motion planning (TAMP) problems remain difficult even with full observability and object-centric states because discrete decisions are tightly coupled to geometr…
- arXiv cs.AIresearchRolling-WAM: World Action Models with Rolling Imagination
arXiv:2609.30247v2 Announce Type: replace-cross Abstract: World Action Models (WAMs) couple action generation with future visual prediction for robotic manipulation. However, completing the joint video-action denoising p…
- arXiv cs.AIagentsRAPID: Robot Agentic Programming from Demonstrations
arXiv:2609.30249v1 Announce Type: cross Abstract: Coding agents have demonstrated enormous success in solving complex programming problems. To leverage their potential for robot systems, this work introduces Robot Agenti…
- arXiv cs.AIagentsLLM Agents Can Easily Tamper With Their Own Traces
arXiv:2609.30266v1 Announce Type: cross Abstract: Asynchronous monitoring, incident investigations, and compliance audits primarily rely on agent traces to reconstruct what happened. These analyses assume that LLM agents…
- arXiv cs.AIresearchGenerating Interesting Scientific Ideas using Knowledge Graphs and LLMs: Evaluations with 100 Research Group Leaders
arXiv:2405.17044v4 Announce Type: replace Abstract: The rapid growth of scientific literature makes it increasingly challenging for researchers to identify novel and impactful ideas, especially across disciplines. Modern…
- arXiv cs.AIresearchUnraveling the cognitive patterns of Large Language Models through module communities
arXiv:2508.18192v2 Announce Type: replace Abstract: Large Language Models (LLMs) have reshaped our world with significant advancements in science, engineering, and society through applications ranging from scientific dis…
- arXiv cs.AIresearchLOGIC: Efficient and Robust Contextual Biasing for Speech LLMs via Logit-Space Integration
arXiv:2601.15397v4 Announce Type: replace Abstract: Recognizing entity phrases remains a critical challenge for speech large language models. Existing prompting methods lack an explicit decoding-time biasing weight, limi…