AI news archive.
19,854 articles, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchMPIDetector: Detecting Multi-modal Prompt Injection via LLM-based Reasoning
arXiv:2610.30202v1 Announce Type: cross Abstract: With the rapid integration of Multi-modal Large Language Models (MLLMs) into real-world applications, Prompt Injection (PI) has emerged as a critical security threat. Att…
- arXiv cs.AIresearchMoEZip: Routing-Aware KV Cache Compression for Sparse Mixture-of-Experts LLMs
arXiv:2610.30206v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) language models scale model capacity efficiently by activating a small subset of parameters per token. However, in long-context inference, memory…
- arXiv cs.AIresearchPrediction Is Not Calibration: Item Parameter Recovery and Representation Coupling in Knowledge Tracing
arXiv:2610.30211v1 Announce Type: cross Abstract: Knowledge tracing (KT) models can estimate item parameters through an item response decoder trained solely for next-response prediction, yet predictive accuracy does not…
- arXiv cs.AIresearchWhen Reconstruction Is Not Enough: Doubly Robust Counterfactual and Interventional Mediation Analysis with Learned Multimodal Representations
arXiv:2610.30220v1 Announce Type: cross Abstract: Learned representations of high-dimensional multimodal health records are increasingly used to adjust for confounding, yet they are usually trained to reconstruct the dat…
- arXiv cs.AIresearchAutomated screening of privacy risks in medical AI
arXiv:2610.30229v1 Announce Type: cross Abstract: Medical artificial intelligence (AI) models are often shared without their training data, assuming this protects the patients who contributed them. However, a model can r…
- arXiv cs.AIresearchEdit, Don't Imitate: Minimal-Edit On-Policy Distillation for Multimodal Reasoning
arXiv:2610.30231v1 Announce Type: cross Abstract: Visual mathematical reasoning can fail before reasoning begins: when a vision-language model (VLM) misreads a diagram, subsequent reasoning proceeds from false premises.…
- arXiv cs.AIagentsSWE-VERIFIER: Selective Verification for Cost-Efficient Coding Agents
arXiv:2610.30284v1 Announce Type: cross Abstract: Coding agents spend substantial computation generating actions and processing long interaction histories. Poor action choices can steer agents into low-value states, wast…
- arXiv cs.AIresearchGeneralizing Metropolis-Hastings for DAG Sampling
arXiv:2610.30291v1 Announce Type: cross Abstract: A fundamental challenge in Bayesian network structure learning lies in efficiently approximating the posterior distribution of directed acyclic graphs (DAGs). For large,…
- arXiv cs.AIagentsGAV: A scientific experimentation environment for PDE agents
arXiv:2610.30292v1 Announce Type: cross Abstract: Reliable feedback is essential for coding agents to develop programs for partial differential equations (PDEs). Yet evaluating these programs on new PDE problems often re…
- arXiv cs.AIresearchGender, Sexual Orientation, and the One-Year Refugee Bar: A Computational Study of Bill C-12 and Canadian Refugee Adjudication
arXiv:2610.30297v1 Announce Type: cross Abstract: In March 2026, Bill C-12 amended the Immigration and Refugee Protection Act to make refugee claims ineligible for referral to the Refugee Protection Division where they a…
- arXiv cs.AIagentsTerraTransfer++: Demonstration-Free End-to-End Planner
arXiv:2610.30319v1 Announce Type: cross Abstract: End-to-end autonomous driving has achieved state-of-the-art performance on benchmarks and real-world deployments. Its standard training recipe, however, is expensive acro…
- arXiv cs.AIresearchMPBRS: Multi-Potential-Based Reward Shaping for Unified Reward Configuration in Humanoid Locomotion
arXiv:2610.30334v1 Announce Type: cross Abstract: Designing and tuning reward functions remains one of the main challenges in building effective reinforcement learning (RL) pipelines for humanoid locomotion. While shapin…
- arXiv cs.AIresearchConsistency Is Not Validity: Self-Rewarded Moral Argumentation Makes Language Models Decisive but Pliable
arXiv:2610.30349v1 Announce Type: cross Abstract: LLMs are increasingly consulted for moral judgment. We aim to improve performance on such tasks by rewarding good arguments, on the premise that arguing well increases th…
- arXiv cs.AIresearchProbing and Steering for Safety-Relevant Behaviours in Large Language Models
arXiv:2610.30363v1 Announce Type: cross Abstract: Activation probes are computed directions in a language model's residual stream, used both to detect safety-relevant behaviours and, by adding their direction back into t…
- arXiv cs.AIagentsMemPlan: Memory-Conditioned PDDL Planning for Partially Observable Text Environments
arXiv:2610.30379v1 Announce Type: cross Abstract: LLM agents are flexible in language interaction, but using them as long-horizon planners is costly and prone to hallucinated or invalid actions. This paper targets a diff…
- arXiv cs.AIresearchAcoustiPrint: Physics-Guided Neural Design of Manufacturable Acoustic Structures
arXiv:2610.30411v1 Announce Type: cross Abstract: Recent work in additive manufacturing and neural inverse design enables the creation of complex three-dimensional structures to manipulate physical fields, yet these desi…
- arXiv cs.AIagentsAgent Mosaic: Recognizing Confidentiality Leaks in Agentic LLM Systems
arXiv:2610.30418v1 Announce Type: cross Abstract: Modern LLM agents can retrieve information from accessible documents, and then disclose a sensitive conclusion assembled from them. This mosaic effect creates a critical…
- arXiv cs.AIresearchHigh-Value Training Data from Noisy, Diverse Sources
arXiv:2610.30421v1 Announce Type: cross Abstract: Training state-of-the-art models increasingly requires learning from large, heterogeneous datasets whose examples vary in quality, domain relevance, and source of supervi…
- arXiv cs.AIagentsSelf-Assessment Reinforcement Learning for Search-State Awareness in Retrieval-Augmented Large Language Models
arXiv:2610.30435v1 Announce Type: cross Abstract: Recent advances in search agents have enabled large language models to tackle complex tasks by retrieving and synthesizing external information. However, existing search…
- arXiv cs.AIresearchaggmetrics: Aggregating Classification Performance Measures for Machine Learning
arXiv:2610.30456v1 Announce Type: cross Abstract: aggmetrics is a Python package that extends any binary performance measure expressed in the four confusion-matrix counts to multiclass, multilabel, ordinal, cost-sensitiv…
- arXiv cs.AIresearchShort-to-Long: Detached Features for Long-Horizon Value Learning
arXiv:2610.30461v1 Announce Type: cross Abstract: Learning useful representations from continually changing value targets remains a challenge in deep reinforcement learning. We study the temporal horizon of this supervis…
- arXiv cs.AIagentsSplatLab: Autonomous Scientific Discovery for 3DGS Research
arXiv:2610.30468v1 Announce Type: cross Abstract: Reproducing a published 3DGS paper takes weeks of expert engineering before any extension can be proposed. Prior autonomous-research systems read papers, draft hypotheses…
- arXiv cs.AIresearchRethinking SAE Evaluation: An Atomic Interpretable Unit Framework Reveals Hidden Polysemanticity and Redundancy
arXiv:2610.30474v2 Announce Type: replace-cross Abstract: Sparse autoencoders (SAEs) were originally motivated by discovering monosemantic features, a property that characterizes individual features but does not specify…
- arXiv cs.AIagentsExternalise the Memory: A Pre-Registered Study of Where an Embodied Agent's Experience Should Live
arXiv:2610.30478v1 Announce Type: cross Abstract: Should an embodied agent's experience be baked into its policy weights, or kept external and retrieved at inference? A memory-augmented teacher (a VLM that retrieves less…
- arXiv cs.AIresearchPhysibleMDM: Enforcing Inference-Time Physical Feasibility Constraints on Motion Diffusion Models via Noise Optimization
arXiv:2610.30500v1 Announce Type: cross Abstract: Motion Diffusion Models (MDMs) can generate diverse and visually appealing human motion; yet, the resulting trajectories frequently violate physical constraints, limiting…
- arXiv cs.AIresearchBehind Aggregate Scores: What Systematically Varying Arena-Hard's LLM Judge Reveals
arXiv:2610.30512v1 Announce Type: cross Abstract: LLM judges are used to determine which of two answers is better. Trusting these verdicts requires knowing how repeatable they are and how much they depend on the judge's…
- arXiv cs.AIresearchProtegrity-PII Technical Report
arXiv:2610.30518v2 Announce Type: replace-cross Abstract: We present Protegrity-PII, a new PII detection model for structured and unstructured text. Protegrity-PII achieves a typed recall of 0.910 with a precision of 0.9…
- arXiv cs.AIresearchLLM2Cert: LLM Synthesis of Certificates with SMT/Lean Verification
arXiv:2610.30524v1 Announce Type: cross Abstract: Lyapunov functions and barrier certificates give state-space guarantees of stability and safety, but synthesizing and verifying them---especially for nonlinear, high-dime…
- arXiv cs.AIagentsDecision Making and Ride Comfort Evaluation Based on Lane-Change Prediction for Autonomous Vehicles
arXiv:2610.30528v1 Announce Type: cross Abstract: Predicting surrounding vehicles' lane-change intentions enables autonomous vehicles to anticipate potential hazards and respond proactively rather than relying solely on…
- arXiv cs.AIagentsLifetime-valid decisions by continually evolving AI analysts
arXiv:2610.30562v1 Announce Type: cross Abstract: AI agents increasingly act as analysts whose decision can often be cast as hypothesis tests. Classical tests control the Type I error rate, the probability of rejecting a…
- arXiv cs.AIresearchBridging the Closed-World Gap: Interpretable Anomaly Detection via a Hybrid TSK-FCM Ensemble on the BETH Dataset
arXiv:2610.30576v1 Announce Type: cross Abstract: A fundamental paradox plagues the deployment of Machine Learning in cybersecurity: supervised models generalize poorly to novel, Out-of-Distribution (OoD) attacks, wherea…
- arXiv cs.AIresearchVisual Grounding First, Multimodal In-context Learning Follows
arXiv:2610.30592v1 Announce Type: cross Abstract: LLMs have shown strong in-context learning (ICL) capability, but extending it to Multimodal LLMs (MLLMs) remains challenging. Prior multimodal ICL methods often rely on I…
- arXiv cs.AIresearchImpacts of generative AI skill development on academic adaptation and learning outcomes among university students
arXiv:2610.30599v1 Announce Type: cross Abstract: Understanding how university students perceive the development of their generative artificial intelligence (GenAI) skills over time can clarify how GenAI use relates to l…
- arXiv cs.AIresearchFuture of Time Series Anomaly Detection: From Historical Review to Unified Outlook
arXiv:2610.30628v1 Announce Type: cross Abstract: Time series anomaly detection (TSAD) is a foundational problem in data mining, aiming to identify observations that deviate significantly from nominal temporal patterns.…
- arXiv cs.AIresearchInterpreting Model Reasoning through Behavioral Counterfactuals
arXiv:2610.30630v1 Announce Type: cross Abstract: Faithfully understanding the internal process that occurs as models form their chain of thought is an important problem with the rise in model capabilities. However, this…
- arXiv cs.AIagentsAsking Less, Adapting Better: Feedback-Efficient Test-Time Adaptation for Vision-and-Language Navigation
arXiv:2610.30643v1 Announce Type: cross Abstract: Vision-and-Language Navigation (VLN) requires agents to follow instructions and navigate to target locations. However, current VLN agents often suffer severe performance…
- arXiv cs.AIresearchOne-shot federated unsupervised domain adaptation with Smoothed Knowledge Distillation and teacher refinement
arXiv:2610.30646v1 Announce Type: cross Abstract: Federated unsupervised domain adaptation (FUDA) addresses the challenge of adapting models to an unlabeled target domain using decentralized source domains while preservi…
- arXiv cs.AIresearchSpectral Smoothing and Spiking Perturbations Reconfigure Capability Profiles and Selective Task Performance in Large Language Models
arXiv:2610.30676v1 Announce Type: cross Abstract: Singular-value spectra encode important structure in transformer feed-forward projections, yet their causal role in downstream behavior remains underexplored. We study po…
- arXiv cs.AIresearchFrom Clinical Judgment to Checkable Rules: Clinician-Led Guardrails for Mental-Health Chatbots
arXiv:2610.30690v1 Announce Type: cross Abstract: Large language models (LLMs) deployed in mental-health contexts can be evaluated for safety failures, but the resulting judgments are often difficult to inspect or explai…
- arXiv cs.AIresearchLoRA Superposition across Pretrained Models
arXiv:2610.30691v1 Announce Type: cross Abstract: Low-rank adaptation (LoRA) is usually tied to the pretrained backbone on which it is learned. We ask whether LoRAs trained independently for the same task on different pr…
- arXiv cs.AIresearchULME: Predictive Coupling for Mutually Reinforcing Embedding and Language Modeling
arXiv:2610.30705v1 Announce Type: cross Abstract: Text embedding and text generation are two of the most useful capabilities of modern large language models, yet they are typically provided by separate specialized models…
- arXiv cs.AIresearchMeaning Contracts: Local Semantic Geometry in Large Language Models
arXiv:2610.30715v1 Announce Type: cross Abstract: This paper studies the local geometry of semantic meaning in large language models. Current representation analyses ask whether meanings are close in embedding space, or…
- arXiv cs.AIresearchLearning from Next-Frame Prediction: Towards Unified and Scalable Visual World Modeling
arXiv:2610.30724v2 Announce Type: replace-cross Abstract: Video provides a natural source of supervision: what happens next serves as a learning signal for visual world modeling. Yet masked video pretraining focuses on r…
- arXiv cs.AIagentsThe Delegation Divide: An Open Index of AI-Agent Delegability across Occupations and Countries
arXiv:2610.30752v1 Announce Type: cross Abstract: 35.2% of task time in world employment is at least partly delegable to AI agents in 2026, and only 2.5% can be handed over in full. Existing measures score exposure to ge…
- arXiv cs.AIagentsAbstract Priorities, Concrete Rewards: How Stated Preferences Relate to What People Reward in AI Responses
arXiv:2610.30763v1 Announce Type: cross Abstract: AI assistants built on large language models (LLMs) are trained on human feedback and increasingly personalized to users' stated preferences. Yet how closely these prefer…
- arXiv cs.AIresearchVLTest: Finding Joint Image-Text Failures in Vision-Language Models
arXiv:2610.30794v1 Announce Type: cross Abstract: Vision-language models (VLMs) are increasingly embedded in software systems. Testing them raises two problems. Mutating either input can change the expected answer, so an…
- arXiv cs.AIresearchNeural Spectral Capacity with Residual Context for Heterogeneous Resource Allocation within Transformers
arXiv:2610.30823v1 Announce Type: cross Abstract: Predicting relative trained performance from architectural specifications can guide design when training costs limit experimentation. Neural Spectral Capacity (NSC) provi…
- arXiv cs.AIresearchOne Evaluation Is Enough: Inversion-Free Image Editing with Trajectory Correction
arXiv:2610.30827v1 Announce Type: cross Abstract: Real-image editing commonly relies on inversion and iterative sampling, while efficient one-step editors often stabilize a large update by estimating and averaging editin…
- arXiv cs.AIresearchBifurcations from Weight Space: JEPAWG on Parametric Dynamical Systems
arXiv:2610.30847v1 Announce Type: cross Abstract: Trained network weights are increasingly treated as data, but there is rarely ground truth about what a population of models should encode, which hampers systematic valid…
- arXiv cs.AIresearchDirect Forcing: Single-Stage Training with Representation-Space Supervision for Few-Step Autoregressive Video Generation
arXiv:2610.30861v2 Announce Type: replace-cross Abstract: Few-step autoregressive video generation enables low-latency streaming, yet prevailing training pipelines separate causal adaptation from distribution matching. R…