AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchSmart Picks in the Dark: Towards Efficient RLVR for Reasoning via Tracing Metacognitive Pivots
arXiv:2606.04503v2 Announce Type: replace-cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has greatly advanced large reasoning models (LRMs), but it requires timely training on a huge fully-annotate…
- arXiv cs.AIresearch'Your AI Text is not Mine': Redefining and Evaluating AI-generated Text Detection under Realistic Assumptions
arXiv:2606.04906v2 Announce Type: replace-cross Abstract: Although it is generally agreed that AI-generated text poses a broad societal risk, there is no common understanding in the AI-generated text detection literature…
- arXiv cs.AIresearchBalancing Image Compression and Generation with Bootstrapped Tokenization
arXiv:2606.05552v2 Announce Type: replace-cross Abstract: Despite progress in image tokenization, standard methods encode redundant information by mixing all granularities within each token, thus redundancy persists betw…
- arXiv cs.AIresearchEnd-to-End Context Compression at Scale
arXiv:2606.09659v2 Announce Type: replace-cross Abstract: Long-context language model inference is bottlenecked by memory, as the KV cache grows with context length. Recent techniques to compress the KV cache fall short:…
- arXiv cs.AIresearchCANVAS: Captioning Art with Narrative Visual-Audio AI Systems
arXiv:2606.09846v2 Announce Type: replace-cross Abstract: For a person who is blind or has low vision (BLV), a painting is usually represented by a few words of alt-text, if anything at all. That text rarely conveys the…
- arXiv cs.AIresearchEDITH: Teaching Robot to Follow Nonverbal Human Intent from Egocentric Vision and Gaze
arXiv:2606.10276v2 Announce Type: replace-cross Abstract: For natural human-robot interaction, a robot must understand human intent expressed not only through language but also through nonverbal signals such as gestures…
- arXiv cs.AIresearchOptimal Post-Training Quantization Scales and Where to Find Them
arXiv:2606.10890v2 Announce Type: replace-cross Abstract: Post-training quantization (PTQ) compresses large language models by mapping weights to low-bit representations. While much work has focused on how weights are ro…
- arXiv cs.AIresearchIs LLM's Moral Decision Compositional?
arXiv:2606.11232v2 Announce Type: replace-cross Abstract: Moral evaluations of language models typically infer a model's values from its choices between single acts, yet in everyday life a person is judged on their histo…
- arXiv cs.AIresearchNon-Gaussianity Reveals Interpretable Directions in Language Models
arXiv:2606.11722v2 Announce Type: replace-cross Abstract: Finding human-interpretable directions in language-model activations is important for understanding and controlling model behavior. These directions are usually d…
- arXiv cs.AIresearchThe Signs Were Always There: Training-Free Detection and Control in Raw Transformer Dimensions
arXiv:2606.12629v4 Announce Type: replace-cross Abstract: Because of polysemanticity and superposition, it has been challenging to attribute token understanding and generation to specific neurons or dimensions. Yet the i…
- arXiv cs.AIresearchAn integrated interpretable control effectiveness learning and nonlinear control allocation methodology for overactuated aircraft
arXiv:2606.13794v3 Announce Type: replace-cross Abstract: Nonlinear dynamics and the strong couplings that arise between multiple effectors undermine the assumptions behind conventional, linear control allocation techniq…
- arXiv cs.AIresearchEvaluating the Robustness of Proof Autoformalization in Lean 4
arXiv:2606.14867v3 Announce Type: replace-cross Abstract: Proof autoformalization aims to translate a mathematical informal proof in natural language into a formal proof in a formal language such as Lean 4. Existing LLM-…
- arXiv cs.AIresearchService-Induced Congestion in Memory-Constrained LLM Serving
arXiv:2606.15555v3 Announce Type: replace-cross Abstract: In large language model (LLM) serving, each request accumulates persistent graphics processing unit (GPU) memory during service as its key-value cache grows with…
- arXiv cs.AIresearchEnd-to-End Radar and Communication Modulation Recognition with Neuromorphic Computing
arXiv:2606.24075v2 Announce Type: replace-cross Abstract: Although deep learning-based methods can achieve high accuracy in automatic modulation recognition (AMR) tasks, their high computational cost makes it difficult t…
- arXiv cs.AIresearchGROW$^2$: Grounding Which and Where for Robot Tool Use
arXiv:2606.30632v2 Announce Type: replace-cross Abstract: Can the robot use a plate to cut a cake if no knife is available? Tool use greatly expands robot capabilities, but to use tools creatively beyond their intended f…
- arXiv cs.AIresearchGemNav: Discrete-Token Visual Robot Navigation using a Multimodal Large Language Model
arXiv:2607.06882v3 Announce Type: replace-cross Abstract: Visual navigation policies built on large pretrained models increasingly rely on large cross-embodiment datasets and expressive action-generation mechanisms. In t…
- arXiv cs.AIresearchFrom Solvers to Research: Large Language Model-Driven Formal Mathematics at the Research Frontier
arXiv:2607.07779v2 Announce Type: replace-cross Abstract: Recent developments in AI for Mathematics (AI4Math), especially Large Language Model (LLM)-driven theorem provers, has achieved remarkable success in formal proof…
- arXiv cs.AIresearchSOS-LoRA: Static Orthogonal-Subspace Low-Rank Adaptation with Fixed Multi-Scale Scaling
arXiv:2607.16252v2 Announce Type: replace-cross Abstract: Low-Rank Adaptation (LoRA) is a widely used parameter-efficient fine-tuning (PEFT) method for large language models. Under a fixed rank budget, LoRA parameterizes…
- arXiv cs.AIresearchArithmetic and its remedies select the evidence on learned geometry
arXiv:2607.17513v2 Announce Type: replace-cross Abstract: Intelligence that must act before outside correction arrives must carry the structure of its expert knowledge, and hyperbolic geometry represents hierarchy with a…
- arXiv cs.AIresearchISO: An RLVR-Native Optimization Stack
arXiv:2607.19331v2 Announce Type: replace-cross Abstract: Reinforcement learning with verifiable rewards (RLVR) advances the reasoning capabilities of language models, yet the optimization layer that converts reward feed…
- arXiv cs.AIresearchNeptuna: A Comprehensive Machine Learning Framework for Benchmarking Complex Multiphase Flows
arXiv:2607.22280v3 Announce Type: replace-cross Abstract: Compressible multiphase flows involving shocks and material interfaces arise in applications such as bubble collapse and droplet breakup, where strong nonlinear i…
- arXiv cs.AIresearchHow Much AI Is in This Track? Quantifying the Proportion of AI-Generated Stems in Hybrid Music Mixtures
arXiv:2608.07285v2 Announce Type: replace-cross Abstract: AI-generated music is increasingly used at the stem level, with producers integrating synthetic drums, basslines, or vocals alongside human-performed instruments.…
- arXiv cs.AIresearchCalling the Bluff: Detecting Ever-Shifting Harmful Chat Dialogue via Ordered Reasoning Chain Regularization
arXiv:2608.08451v2 Announce Type: replace-cross Abstract: Harmful chat dialogues are ever-shifting through type-shifting and lexical evasion, yet we find they share invariant principles, i.e., an Ordered Reasoning Chain…
- arXiv cs.AIresearchTest-Time Augmentation for LLMs: Where to Inject Diversity at Inference Time
arXiv:2608.09351v2 Announce Type: replace-cross Abstract: Aggregating several LLM responses to the same question can improve accuracy, but every method that does so must decide where the responses should differ. Establis…
- arXiv cs.AIresearchSelection, Recombination, or a Fresh Solve? A Candidate-Free Control for Single-Pass Test-Time Aggregation
arXiv:2608.18379v2 Announce Type: replace-cross Abstract: An aggregation pass gives a language model two things: candidate solutions and another chance to answer. We ask whether the candidates improve that attempt, and w…
- arXiv cs.AIresearchEditPPT: Structured Execution and Evaluation for Reliable Slide Editing
arXiv:2608.20381v2 Announce Type: replace-cross Abstract: Reliable slide editing requires precise modifications to intended objects while preserving the rest of the presentation. We argue that slide editing is a problem…
- arXiv cs.AIresearchBeyond Uniform Local Isometry and Topology: FactoMap for Disentangled Representations
arXiv:2608.24762v2 Announce Type: replace-cross Abstract: Many disentanglement methods represent generative factors using Euclidean product coordinates, although the underlying factor spaces may wrap, collapse, or have p…
- arXiv cs.AIresearchMathAdv: What Theorem Provers Know, Reason, Formalize, and Generalize
arXiv:2608.25449v3 Announce Type: replace-cross Abstract: Formal theorem proving enables machine-verifiable evaluation of mathematical reasoning, yet existing benchmarks often emphasize aggregate proof accuracy, concentr…
- arXiv cs.AIresearchGood Memory Has ECC: Evaluating the Memory of Vision-Language Models Beyond Accuracy
arXiv:2609.00103v2 Announce Type: replace-cross Abstract: Memory is widely viewed as an important unsolved problem for LLMs and VLMs, and current benchmarks typically evaluate it by testing accuracy over long text or vid…
- arXiv cs.AIresearchCleaner Speech, Weaker Generalization: Revisiting Pitt-Derived Benchmarks for Alzheimer's Disease Detection
arXiv:2609.00276v2 Announce Type: replace-cross Abstract: Speech-based Alzheimer's disease (AD) detection increasingly relies on speech-enhanced and curated versions of the Pitt Corpus, where speech enhancement, sample s…
- arXiv cs.AIresearchGraFT: A Training-Free Framework for Spatial Reasoning in Multimodal Large Language Models via 3D Scene Graphs
arXiv:2609.03892v2 Announce Type: replace-cross Abstract: 3D spatial reasoning underpins understanding and acting in the physical world, yet it remains unreliable in current multimodal large language models (MLLMs). Thes…
- arXiv cs.AIresearchLeveraging Low-Level Symbolic Competences for Unsupervised Grounding in Hallucination Detection
arXiv:2609.05025v2 Announce Type: replace-cross Abstract: Hallucination-where a language model generates outputs that are factually incorrect or unsupported by the source-is a major challenge for both prompted and fine-t…
- arXiv cs.AIresearchJoint Deployment and Resource Management for Group-Tag-Based Physical-Layer Authentication and Collaborative Communication in NTN
arXiv:2609.09475v2 Announce Type: replace-cross Abstract: Physical-layer authentication (PLA) in non-terrestrial networks (NTNs) has primarily been studied from a security perspective, while its system-level integration…
- arXiv cs.AIresearchChopthin-Consensus Power Sampling: A Diversity-Preserving Approach to LLM Decoding
arXiv:2609.12243v2 Announce Type: replace-cross Abstract: Inference-time power sampling via Sequential Monte Carlo (SMC) can substantially improve large language model (LLM) reasoning without requiring post-training. How…
- arXiv cs.AIresearchDiagnosing Faults in Reinforcement Learning Simulators with Canonical Polynomial Invariants
arXiv:2609.13194v2 Announce Type: replace-cross Abstract: A large literature incorporates physical structure into learned dynamics on the premise that models respecting the underlying physics should predict better. We te…
- arXiv cs.AIresearchBeyond density snapshots: Dynamical memory in Chern-Simons reservoirs
arXiv:2609.13315v2 Announce Type: replace-cross Abstract: How a reservoir encodes input history jointly in density and geometric degrees of freedom remains insufficiently understood. We address this mechanistic gap with…
- NVIDIA Developer BlogresearchDesign Proteins with RFD4-Proteina Using Reward-Guided Inference-Time Search
Protein design is a search problem. Choose amino acids for an eight-residue protein sequence and more than 25 billion combinations are possible. Choose amino...
- LessWrongresearchHuman vs AI in Go in 2026: Lessons for superintelligence?
In July 2026, Korean Go player Shin Jin-seo defeated KataGo 2-1 , with only [1] a 2-stone handicap (favoring Shin). I thought this was very surprising, and posted on LessWrong and Twitter asking how to interpret this. I…
- Guardian AIresearchNew AI tool can predict cancer survival by a single patient selfie, study suggests
System can estimate a person’s biological age and predict chances of cancer survival from a single photograph Doctors have developed an artificial intelligence tool that can predict a cancer patient’s chances of survival…
- LessWrongresearchGLM-5.1 Has ~Astra Level CoT Control
TL;DR. We run the CoT-Control evaluation suite on GLM, Kimi and DeepSeek models and find that while Kimi and DeepSeek models have poor CoT control, GLM models are outliers in their ability to control CoT. We extensively…
- LessWrongresearchGradient Routing Localizes and Removes Unintended Behaviors in RL
Authors: Shawn Hu*, Jake Ward*, Aria Wong, Nathan Hu, Bryce Woodworth, Alexander Matt Turner, Alex Cloud (*Equal contribution, randomly ordered) tl;dr. During training, models sometimes learn to cheat at tasks rather tha…
- LessWrongresearchAI as a Normal Technology should be a vision
"AI as a Normal Technology" (AINT), proposed as a counterweight to "AI as an Abnormal Technology" (AIAbT) makes much more sense as a prescriptive / normative framework than a descriptive one. To be clear, the fact that t…
- MIT Technology Review AIresearch“We were looking at it wrong”: The explosive claim that’s upended the AI world
“I think I have good news about the nature of neural networks.” With those simple words of optimism, Nick Cammarata, a researcher at the AI startup Goodfire, kicked off a 26,000-word blog post that has knocked the AI wor…
- LessWrongresearchTwitch Plays: AI Interpretability of Bonsai-8B
Quick Links Watch the Twitch stream . Check the status of the interpretability map. Do your own parallel research and submit pull requests to fill out the map. What is this? Let's Play Interpretability is a Twitch livest…
- NVIDIA Developer BlogresearchAccelerating Mamba-2 with High-Performance Forward Pass Kernels in NVIDIA CUDA Tile
NVIDIA CUDA Tile (cuTile) is a Python-based DSL and parallel programming model for NVIDIA GPUs that automatically leverages advanced hardware capabilities to...
- AWS Machine Learning BlogresearchPush-based KV transfer to improve TTFT and TPOT for disaggregated inference in vLLM
vLLM's prefill-decode disaggregation splits LLM inference across engine instances but must transfer the KV cache between them. This post introduces a push-based NIXL connector that moves transfer coordination off the cri…
- AWS Machine Learning BlogresearchImprove TTFT for multi-turn agentic workloads with bi-directional KV transfer in vLLM
Multi-turn agentic workloads reprocess growing conversation context each turn, driving up latency. See how bi-directional KV cache transfer lets prefill and decode nodes reuse previously computed context across turns, cu…
- TechRadar AIresearchAI keeps breaking old encryption mysteries — and is dunking on humans too
Mainstream AI news from a large tech publication.
- IEEE Spectrum AIresearchEdge AI for Smart Cities: From Insight to Cognitive Action
This white paper explains how edge AI and physical AI help cities move beyond passive monitoring. It shows how urban infrastructure can sense, reason, and act on real-time conditions across public safety, mobility, and i…
- MIT Technology Review AIresearchAI researchers stranded Claude on a virtual desert island. It started to “play.”
What do AI agents do if you don’t tell them to do anything? The answer looks a lot like play. Researchers repeatedly stranded them alone on a virtual desert island for 30 hours in an experiment reminiscent of the movies…