AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchEPIC: Efficient and Parallel Inference under CFG Constraints for Diffusion Language Models
arXiv:2606.00722v2 Announce Type: replace-cross Abstract: Controlling language model outputs is essential for ensuring structural validity, reliability, and downstream usability, and diffusion language models are no exce…
- arXiv cs.AIresearchLow-Resource Safety Failures Are Action Failures, Not Representation Failures
arXiv:2606.01196v2 Announce Type: replace-cross Abstract: Language models often answer harmful requests in low-resource languages (LRLs) that they refuse in high-resource languages (HRLs). Across three instruction-tuned…
- arXiv cs.AIresearchAll you need to break LLMs are Black-Box, Adapting, Efficient, Transferable, Harmful, Applicable ... Attacks
arXiv:2606.03647v2 Announce Type: replace-cross Abstract: Accurately evaluating adversarial robustness is a longstanding challenge. A flawed attack design can inflate robustness estimates, making deployment risk assessme…
- arXiv cs.AIresearchA Goal-Set Characterization of Task Composition in the Boolean Task Algebra
arXiv:2606.04053v2 Announce Type: replace-cross Abstract: The Boolean Task Algebra (BTA) provides a principled framework for zero-shot task composition in reinforcement learning by equipping goal-reaching tasks with Bool…
- arXiv cs.AIresearchBenchmarks in Leipzig
arXiv:2606.05818v2 Announce Type: replace-cross Abstract: Between April 1 and May 15, 2026, a group of 49 mathematicians compiled a dataset of research-level mathematics questions with known answers. Most of the work was…
- arXiv cs.AIresearchLatentWave: JEPA Pretraining for Wireless Foundation Models
arXiv:2606.06373v2 Announce Type: replace-cross Abstract: Wireless foundation models have emerged as a promising alternative to building separate models for each wireless task. However, existing approaches rely on masked…
- arXiv cs.AIresearchSTREAM: Stochastic Riemannian Flow Matching with Anisotropic Decoder for Digital Histopathology Image Generation
arXiv:2606.07036v3 Announce Type: replace-cross Abstract: Synthetic histopathology image generation addresses patient-privacy concerns and the growing data demands of foundation models. Existing state-of-the-art histopat…
- arXiv cs.AIresearchCultureScore: Evaluating Cultural Faithfulness in Video Generation Models
arXiv:2606.07311v3 Announce Type: replace-cross Abstract: As video generation models like Veo 3.1 and LTX-2 advance, their ability to accurately represent diverse global cultures remains a critical yet understudied front…
- arXiv cs.AIresearchLEAF: Growing Trees Without Branching for Speech-Aware Large Language Model Post-Training
arXiv:2606.07610v2 Announce Type: replace-cross Abstract: State-of-the-art GRPO-style methods for speech-aware large language model post-training suffer from coarse credit assignment, broadcasting the same terminal-rewar…
- arXiv cs.AIresearchGefen: Optimized Stochastic Optimizer
arXiv:2606.13894v4 Announce Type: replace-cross Abstract: AdamW is a default optimizer for deep learning, but its moment states add two parameter-sized buffers to training memory, increasing the cost of large-scale pretr…
- arXiv cs.AIresearchMachine learning modeling of hit-into-play probability and quantification of pitch sequence effects
arXiv:2606.17345v2 Announce Type: replace-cross Abstract: Although pitch sequencing is a central topic in baseball analytics, previous studies have primarily focused on improving predictive performance for the final pitc…
- arXiv cs.AIresearchEG-VQA: Benchmarking Verifiable Video Question Answering with Grounded Temporal Evidence
arXiv:2606.24797v2 Announce Type: replace-cross Abstract: Recent advances in Video Large Language Models (Video-LLMs) have yielded promising performance on Video Question Answering (VideoQA). Nevertheless, existing bench…
- arXiv cs.AIresearchNarrative-UFET: Narrative Generation for Ultra-Fine Entity Typing
arXiv:2606.27598v2 Announce Type: replace-cross Abstract: Ultra-fine entity typing (UFET) assigns highly specific types to entity mentions, but current approaches struggle with types in the long tail. We hypothesize that…
- arXiv cs.AIresearchBehavior Uncloning: Distilling Mode Redirection into Policy Weights without Inference-Time Steering
arXiv:2606.29201v3 Announce Type: replace-cross Abstract: Behavior-cloned policies often learn multiple behavior modes from demonstration datasets, including modes that are unsafe or otherwise undesired at deployment. Fo…
- arXiv cs.AIresearchMulti-Level Distributional Entropy from Flow Summary Statistics for Explainable Network Intrusion Detection
arXiv:2606.29797v2 Announce Type: replace-cross Abstract: Machine learning network intrusion detection systems (IDS) operate on aggregate flow statistics that discard the distributional structure of traffic, and although…
- arXiv cs.AIresearchAnchored or Drifting: What Recursive Self-Generation Reveals About Training Data
arXiv:2606.31991v2 Announce Type: replace-cross Abstract: Large generative models are known to memorize their training data, posing severe privacy risks. Yet, current methods to detect training membership typically rely…
- arXiv cs.AIresearchMeta-Transfer Learning for mmWave Beam Alignment
arXiv:2607.00860v2 Announce Type: replace-cross Abstract: Millimeter-wave (mmWave) beam alignment is critical for next-generation wireless systems, but existing approaches either update all parameters during adaptation o…
- arXiv cs.AIresearchOffice Comprehension Benchmark
arXiv:2607.01245v2 Announce Type: replace-cross Abstract: We introduce Office Comprehension Bench (OCB), the first public benchmark to jointly evaluate LLM systems on Word, Excel, and PowerPoint comprehension over native…
- arXiv cs.AIresearchAnyGroundBench: A Multi-Domain Adaptation Benchmark for Video Grounding in VLMs
arXiv:2607.02269v2 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) have shown strong performance in Spatio-Temporal Video Grounding (STVG), yet they are still evaluated mostly in a zero-shot manner o…
- arXiv cs.AIresearchSafe Inference-Time Alignment via Lagrangian Reward Augmentation
arXiv:2607.02781v2 Announce Type: replace-cross Abstract: Inference-time alignment steers a frozen language model during decoding using auxiliary reward signals, avoiding the cost of repeated weight updates. However, exi…
- arXiv cs.AIresearchMask2Real-WM: Controllable Dexterous World Models via Segmentation Masks as a Sim-to-Real Bridge
arXiv:2607.04546v3 Announce Type: replace-cross Abstract: Learning action-conditioned world models for dexterous manipulation that are genuinely controllable requires capturing complex, high-dimensional hand kinematics f…
- arXiv cs.AIresearchUI2App: Benchmarking Visual Interaction Inference in Executable Web Application Generation
arXiv:2607.06306v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have demonstrated growing competence in generating web pages from UI screenshots, which convey both visual structure and cues to appl…
- arXiv cs.AIresearchCo-LMLM: Continuous-Query Limited Memory Language Models
arXiv:2607.07707v2 Announce Type: replace-cross Abstract: Externalizing knowledge in LLM pre-training is a promising avenue to achieve higher performance at smaller scales, control knowledge use, and overall increase mod…
- arXiv cs.AIresearchAuthority-Bound Governance of Heterogeneous AI Security Decisions in Telecom and IoT Networks
arXiv:2607.09259v2 Announce Type: replace-cross Abstract: Artificial intelligence (AI)-enabled security decision systems in telecom and IoT networks can draw on heterogeneous models whose outputs may trigger operational…
- arXiv cs.AIresearchExact and Certified Data Shapley for Weighted Nearest Neighbor Regression
arXiv:2607.11956v3 Announce Type: replace-cross Abstract: Data Shapley can be computed exactly and efficiently for unweighted K-nearest-neighbor (KNN) models, the basis of the popular KNN-Shapley. For weighted KNN, howev…
- arXiv cs.AIresearchContinual Learning with Elastic Regularization and Synthetic Replay for Federated MLLM Fine-Tuning
arXiv:2607.12112v2 Announce Type: replace-cross Abstract: Federated fine-tuning of Multimodal Large Language Models (MLLMs) across distributed networks enables privacy-sensitive adaptation to evolving data streams, yet a…
- arXiv cs.AIresearchCAS I: A Geometric Coding Theorem
arXiv:2607.13796v2 Announce Type: replace-cross Abstract: This paper establishes a direct analogue of the classical Coding Theorem in the setting of symmetry groups. We consider computable bijections on the set of binary…
- arXiv cs.AIresearchSLPO: Scaling Latent Reasoning via a Surrogate Policy
arXiv:2607.19691v3 Announce Type: replace-cross Abstract: Reinforcement learning with verifiable rewards has become the predominant recipe for eliciting test-time scaling in explicit Chain-of-Thought reasoners. Yet this…
- arXiv cs.AIresearchMedical foundation models converge less under label supervision
arXiv:2607.20274v3 Announce Type: replace-cross Abstract: Diagnostic classifiers and imaging biomarkers are fitted on the embeddings of medical foundation models. These models are replaced as new versions appear. This pr…
- arXiv cs.AIresearchWhen Synthetic Users Fail: A Cross-Domain Benchmark of LLM-Simulated Human Survey Responses
arXiv:2607.26348v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used as synthetic users, stand-ins for human respondents whose simulated answers feed product, policy, and market de…
- arXiv cs.AIresearchCost-Effective Automated Judging of Natural-Language Mathematical Proofs
arXiv:2608.00004v2 Announce Type: replace-cross Abstract: Grading natural-language mathematical proofs is a recurring cost in evaluating math-reasoning systems, and frontier LLM judges are expensive. We ask whether cheap…
- arXiv cs.AIresearchTraining nGPT
arXiv:2608.01284v3 Announce Type: replace-cross Abstract: The normalized Transformer (nGPT) realizes hyperspherical representation learning by constraining model parameter vectors and activation vectors to the unit hyper…
- arXiv cs.AIresearchCLEARMIND: Closed-Form Laplacian Embeddings of Arakelov-Green Resistance for Morphological Intrinsic Neuronal Distances
arXiv:2608.04460v2 Announce Type: replace-cross Abstract: Tropical geometry turns a metric graph into a flat torus, its tropical Jacobian, and measures distances between the images of its points by a closest vector probl…
- arXiv cs.AIresearchSite Is Decodable Before Pretraining: Negative Controls for Probing Frozen Brain-MRI Foundation Models
arXiv:2608.10295v2 Announce Type: replace-cross Abstract: Brain foundation models are often tested with a probe. The model is frozen, a simple classifier is trained on its output, and the classifier's accuracy is taken a…
- arXiv cs.AIresearchThe AI Accountability Ecosystem in the Era of Language Models
arXiv:2608.12320v2 Announce Type: replace-cross Abstract: This article reviews and updates the framework for accountability in AI based on account- ability ecosystems. We update the framework in light of the latest devel…
- arXiv cs.AIresearchLearn What's Left, Not What's Mastered: Saturation Aware Advantage Reweighting for Multi-Reward Policy Optimization
arXiv:2608.16072v2 Announce Type: replace-cross Abstract: Reinforcement learning (RL) with group-relative advantages has become the de facto standard for post-training language model reasoners. However, when optimizing m…
- arXiv cs.AIresearchThe Acknowledgment Point Is the System: Durable Policy-Decision Receipts for AI Audit Evidence
arXiv:2608.17176v2 Announce Type: replace-cross Abstract: An AI audit record is useful only if its durability and trust boundary are explicit. Returning a guarded decision before any durable write minimizes latency, but…
- arXiv cs.AIresearchTRACE: Training-time Report-guided and Clinically Ordered Concept Editing
arXiv:2608.20809v2 Announce Type: replace-cross Abstract: Breast ultrasound diagnosis relies on clinically meaningful semantic concepts, yet most deep learning methods adopt end-to-end image-to-label paradigms that lack…
- arXiv cs.AIresearchDoes Fine-Tuning Undo Activation Steering? Behavioural Recovery Without Weight-Edit Reversal
arXiv:2608.24988v2 Announce Type: replace-cross Abstract: Activation steering can be embedded directly into a language model's weights, shaping behaviour without inference-time intervention and offering a way to encode a…
- arXiv cs.AIresearchPointing the Way, Hiding the Destination: Practical Private Dense Retrieval at Scale
arXiv:2608.25735v2 Announce Type: replace-cross Abstract: Hosted retrieval-augmented generation (RAG) and semantic search allow users to query valuable provider-held corpora, raising two competing demands: to hide each q…
- arXiv cs.AIresearchDifference-in-Differences on a Censored Rating Scale Can Manufacture an Effect: Evidence from a Pre-Registered LLM-Judge Audit
arXiv:2608.27309v5 Announce Type: replace-cross Abstract: LLM-judge audits assess bias by comparing ratings across matched conditions. Difference-in-differences designs take the rating difference between two candidate re…
- arXiv cs.AIresearchThe Hallucination Signal Is a Mean Shift: Why Simple Probes Suffice
arXiv:2608.28930v2 Announce Type: replace-cross Abstract: Hidden-state probes effectively detect LLM hallucinations, but the geometry of the signal remains poorly characterized, driving increasingly complex probe archite…
- arXiv cs.AIresearchCPR for LLMs: Critical-Point Routing against Catastrophic Forgetting in Domain Adaptation
arXiv:2608.30158v4 Announce Type: replace-cross Abstract: Supervised fine-tuning (SFT) is the de facto standard for adapting large language models (LLMs) to target domains, but it often degrades the model's general capab…
- arXiv cs.AIresearchCapability-Gated Language Models: Security Composes, Utility Does Not
arXiv:2609.00445v3 Announce Type: replace-cross Abstract: Deployed language model safeguards (safety fine-tuning, filtering, unlearning) vary by principal only outside the model weights: filters are reconfigured, tiers a…
- arXiv cs.AIresearchA Measured Communication-Quality Frontier for Federated LoRA Fine-Tuning with Adaptive Phase-Switching
arXiv:2609.13512v2 Announce Type: replace-cross Abstract: Federated fine-tuning of large language models with low-rank adaptation (LoRA) reduces the number of trainable parameters, but communication remains the dominant…
- arXiv cs.AIresearchTriCalRAG: A Three-Strategy, Retrieval-Augmented Benchmark for On-Premise LLM-Based Root Cause Analysis in AIOps
arXiv:2609.14762v2 Announce Type: replace-cross Abstract: Operational logs create a need for private, resource-efficient incident analysis, but aggregate detection scores can conceal severe prediction bias. We present Tr…
- LessWrongresearchTakes One to Know One - Training a model to grade reward hacks causes it to reward hack less itself
Code The question I asked in this project was - if I finetune a model to judge/catch reward hacks, does it change its behavior when completing the tests itself. Does it learn to hack more or less? Why it’s worth pursuing…
- LessWrongresearchYou can graft SDF changes from base models onto post-trained models
Webpage: praxis-research.org/grafting | Paper link: https://arxiv.org/abs/2610.00767 Work done during Peter and Dani's MATS Fellowship (10.0) under the mentorship of Shi, with Jinghua Ou as Research Manager. TLDR; We sha…
- Apple Machine Learning ResearchresearchRISED: Rubrics for Agentic Multi-Environment Selection and Self-Distillation
Training a single LLM agent jointly across diverse interactive environments has attracted increasing attention as a route to generalist agents. Existing curriculum and data-selection strategies often allocate training at…
- The Verge AIresearchAll the drama around AI’s takeover of mathematics
This past year, OpenAI, Anthropic, and other labs have announced breakthroughs on numerous long-standing mathematical problems, in some cases pushing well beyond what researchers expected current systems to be capable of…