AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchPPS: Plug-and-Play Saccadic Vision for Fine-Grained Classification
arXiv:2509.15688v2 Announce Type: replace-cross Abstract: Fine-grained visual classification benefits from high-resolution inputs to resolve subtle, localized cues, but at the cost of large compute and memory budgets. In…
- arXiv cs.AIresearchMLLMEraser: Achieving Test-Time Unlearning in Multimodal Large Language Models through Activation Steering
arXiv:2510.04217v4 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) have demonstrated remarkable capabilities across vision-language tasks, yet their large-scale deployment raises pressing…
- arXiv cs.AIresearchOn the Faithfulness of Visual Thinking: Measurement and Enhancement
arXiv:2510.23482v2 Announce Type: replace-cross Abstract: Recent large vision-language models (LVLMs) can generate vision-text multimodal chain-of-thought (MCoT) traces after reinforcement fine-tuning (RFT). However, we…
- arXiv cs.AIresearchSpeech-DRAME: A Framework for Human-Aligned Benchmarks in Speech Role-Play
arXiv:2511.01261v2 Announce Type: replace-cross Abstract: Evaluating spoken role-play requires judging not only linguistic content but also prosody, emotional expression, character embodiment, and contextual fit. Existin…
- arXiv cs.AIresearchDiTGaze: Geometry-Conditioned Diffusion Augmentation for Few-Shot Gaze Estimation
arXiv:2511.11231v2 Announce Type: replace-cross Abstract: Accurate person-specific gaze estimation benefits from labeled user data, but collecting images that densely cover gaze directions is burdensome. We present DiTGa…
- arXiv cs.AIresearchInstruction-Tuning Open-Weight Language Models for BPMN Model Generation
arXiv:2512.12063v2 Announce Type: replace-cross Abstract: Domain models are central to software engineering, as they enable a shared understanding, guide implementation, and support automated analyses and model-driven de…
- arXiv cs.AIresearchScaling Causal Mediation for Complex Systems: A Framework for Root Cause Analysis
arXiv:2512.14764v2 Announce Type: replace-cross Abstract: Modern operational systems ranging from logistics and cloud infrastructure to industrial IoT, are governed by complex, interdependent processes. Understanding how…
- arXiv cs.AIresearchVariational decomposition autoencoding improves disentanglement of latent representations
arXiv:2601.06844v2 Announce Type: replace-cross Abstract: Understanding the structure of complex, nonstationary, high-dimensional time-evolving signals is a central challenge in scientific data analysis. In many domains,…
- arXiv cs.AIresearchTIRx: A Unified Tile Primitive ML Compiler Across Heterogeneous Backends
arXiv:2601.19092v3 Announce Type: replace-cross Abstract: Modern ML kernels evolve rapidly with new optimizations and hardware features. ML compiler abstractions must accommodate richer memory hierarchies, more forms of…
- arXiv cs.AIresearchEroding the Truth-Default: A Causal Analysis of Human Susceptibility to Foundation Model Hallucinations and Disinformation in the Wild
arXiv:2601.22871v3 Announce Type: replace-cross Abstract: As foundation models (FMs) approach human-level fluency, distinguishing synthetic from organic content has become a key challenge for Trustworthy Web Intelligence…
- arXiv cs.AIresearchSafe Reinforcement Learning via Recovery-based Shielding with Gaussian Process Dynamics Models
arXiv:2602.12444v3 Announce Type: replace-cross Abstract: Reinforcement learning (RL) is a powerful framework for optimal decision-making and control but often lacks provable guarantees for safety-critical applications.…
- arXiv cs.AIresearchInterpretability-by-Design with Accurate Locally Additive Models and Conditional Feature Effects
arXiv:2602.16503v2 Announce Type: replace-cross Abstract: Generalized additive models (GAMs) offer high interpretability through independent univariate feature effects but underfit when interactions are present in data.…
- arXiv cs.AIresearchSUREON: A Benchmark and Vision-Language Model for Surgical Reasoning
arXiv:2603.06570v2 Announce Type: replace-cross Abstract: A capable surgical AI model should reason from visual evidence; yet, many existing surgical benchmarks can be partially solved without visual input, as models may…
- arXiv cs.AIresearchProof-Carrying Materials: Falsifiable Safety Certificates for Machine-Learned Interatomic Potentials
arXiv:2603.12183v3 Announce Type: replace-cross Abstract: Machine-learned interatomic potentials (MLIPs) are deployed for high-throughput materials screening without formal reliability guarantees. We show that a single M…
- arXiv cs.AIresearchSpecification-Aware Distribution Shaping for Robotics Foundation Models
arXiv:2603.17969v2 Announce Type: replace-cross Abstract: Robotics foundation models have demonstrated strong capabilities in executing natural language instructions across diverse tasks and environments. However, they r…
- arXiv cs.AIresearchEnhancing Safety of Large Language Models via Embedding Space Separation
arXiv:2603.20206v2 Announce Type: replace-cross Abstract: Safety alignment does not fully prevent large language models (LLMs) from producing harmful content under adversarial attacks. We observe that existing attack met…
- arXiv cs.AIresearchCan Humans Tell? A Dual-Axis Study of Human Perception of LLM-Generated News
arXiv:2604.03755v2 Announce Type: replace-cross Abstract: Can humans tell whether a news article was written by a person or a large language model (LLM)? We investigate this question using JudgeGPT, a study platform that…
- arXiv cs.AIresearchAutoregressive Diffusion for Compositional Text-to-Scene Generation
arXiv:2604.16552v3 Announce Type: replace-cross Abstract: Recent text-to-scene generation approaches largely reduced the manual efforts required to create 3D scenes. However, their focus is either to generate a scene lay…
- arXiv cs.AIresearchHow Fast Should a Model Commit to Supervision? Training Reasoning Models on the Tsallis Loss Continuum
arXiv:2604.25907v3 Announce Type: replace-cross Abstract: The standard recipe for training reasoning models---SFT on rationale annotations followed by RLVR---relies on rationale annotations being available and clean. Whe…
- arXiv cs.AIresearchAMMA: A Multi-Chiplet Memory-Centric Architecture for Low-Latency 1M Context Attention Serving
arXiv:2604.26103v3 Announce Type: replace-cross Abstract: All current LLM serving systems place the GPU at the center, from production-level attention-FFN disaggregation to NVIDIA's Rubin GPU-LPU heterogeneous platform.…
- arXiv cs.AIresearchSparkle: Realizing Lively Instruction-Guided Video Background Replacement via Decoupled Guidance
arXiv:2605.06535v2 Announce Type: replace-cross Abstract: In recent years, open-source efforts like Senorita-2M have propelled video editing toward natural language instruction. However, current publicly available datase…
- arXiv cs.AIresearchDRScaffold: Boosting Dense-Scene Reasoning in Lightweight Vision Language Models
arXiv:2605.26038v2 Announce Type: replace-cross Abstract: Lightweight vision-language models perform competitively on standard benchmarks yet fail systematically in dense-scene reasoning, where multiple objects, attribut…
- arXiv cs.AIresearchAnalyzing the Optimization Conflict Between Image- and Text-Based Person Re-Identification
arXiv:2606.02242v2 Announce Type: replace-cross Abstract: Unifying image-based (I2I) and text-based (T2I) person re-identification (ReID) within a single model is complicated by a fundamental conflict between their train…
- arXiv cs.AIresearchHow Do Video Foundation Models Encode Intuitive Physics? Probing Across Pretraining Paradigms
arXiv:2606.09646v3 Announce Type: replace-cross Abstract: Do video foundation models encode a usable sense of how the physical world works, and does intuitive physics emerge naturally from large-scale video pretraining?…
- arXiv cs.AIresearchBeyond IID: How General Are Tabular Foundation Models, Really?
arXiv:2606.30410v2 Announce Type: replace-cross Abstract: Foundation models for predictive machine learning on tabular data have recently gained significant traction in academia and industry. Research communities across…
- arXiv cs.AIresearchTowards Diverse and Comprehensive Benchmarks for Mutual Information Estimation
arXiv:2607.03487v2 Announce Type: replace-cross Abstract: Mutual information estimation is a central problem in machine learning and statistics; however, existing benchmarks typically evaluate estimators on simplified, l…
- arXiv cs.AIresearchDAUPNet: Domain-Aware Uncertainty Modeling for Reliable Prototype Discrimination in Cross-Domain Few-Shot Semantic Segmentation
arXiv:2607.16308v2 Announce Type: replace-cross Abstract: Cross-domain few-shot semantic segmentation (CD-FSS) has predominantly been formulated as learning domain-invariant representations or improving support-query cor…
- arXiv cs.AIresearchFree energy landscape of Dense Associative Memory
arXiv:2607.19195v3 Announce Type: replace-cross Abstract: Using large deviations theory, we solve and obtain a general expression for the free energy functional for a broad class of associative memories, including dense…
- arXiv cs.AIresearchCogBattery: Human-Referenced Evaluation of Cognitive Task Behavior and Structure in Language Models
arXiv:2607.24999v2 Announce Type: replace-cross Abstract: Human cognitive tasks let us compare language models with humans under controlled conditions. Yet a model can reproduce the direction of a human effect while maki…
- arXiv cs.AIresearchReBRAC-v2: The Return of the King
arXiv:2608.01205v2 Announce Type: replace-cross Abstract: Recent offline reinforcement learning methods increasingly rely on expressive generative policies and specialized value guidance. We ask how far systematic modern…
- arXiv cs.AIresearchSimulation-free Unbalanced Dynamic Optimal Transport with General Growth Penalty
arXiv:2609.04710v2 Announce Type: replace-cross Abstract: Inferring cellular dynamics from unpaired single-cell snapshots requires modeling both state transitions and population growth or death. Unbalanced dynamic optima…
- arXiv cs.AIresearchHyCO: A Hybrid Neural Solver for Combinatorial Optimization
arXiv:2609.07990v2 Announce Type: replace-cross Abstract: Sequential reinforcement learning (RL) solvers and global diffusion model (DM) solvers for neural combinatorial optimization exhibit complementary failure modes u…
- LessWrongresearchAnthropomorphic vs Biological reasoning for thinking about human analogies for LLM behavior
TL;DR I try to distinguish between what I call ‘anthropomorphic’ vs ‘biological’ reasoning and look at different ways in which they are used. Note: This post is a small part of a larger set of ideas that makes more speci…
- LessWrongresearchIs your chatbot less helpful because it thinks you’re a bad person? — Replicating a blogpost
TL;DR [1] A blogpost from a few months ago suggested that LLMs might provide less helpful answers if they see the user as a bad person. I replicate the experiment on more recent models: Opus 5.5, Sonnet 5.5, Haiku 5.5, G…
- The DecoderresearchAdobe is testing how well AI agents can use editing tools like a Photoshop professional would
Adobe Research and UC San Diego built EditAgent, an AI system that edits images more like a Photoshop pro. Instead of retrying the same general-purpose editor over and over, the agent picks the right specialized tool for…
- The DecoderresearchChinese AI models are politically complex, not just government mouthpieces, study finds
A study by Vals AI tested 25 AI models for alignment with Chinese government positions and found surprising results. Chinese models respond to political questions less uniformly than expected, and Western models occasion…
- LessWrongresearchWhat would you do if you had all the time in the world?
Suppose I tell you that next week three things will happen: A new science paper is published definitively proving that DNN scaling hits heavy diminishing returns at around 2x the current level of compute Major labs are r…
- LessWrongresearchLegible Networks: recasting neural nets so you can read them exactly [linkpost]
I've made some interesting strides in AI interpretability: recasting trained neural nets into a geometry where you can read them exactly, with arithmetic — the exact counterpart of Anthropic's averaged "J-lens," includin…
- SiliconANGLE AIresearchCan enterprise AI pay for the boom?
The artificial intelligence buildout is making an enormous bet on what businesses will eventually be willing to pay for intelligence. In our research this summer on forecasting the AI bubble, we argued that a key indicat…
- LessWrongresearchGeneralized belief states in transformers
Many thanks to Xavier Poncini, Paul Riechers, Fernando Rosas, and Matt Smith for useful conversations. 1. Introduction Despite the recent successes of transformers and other sequence models, there are still many open que…
- LessWrongresearchDialogue with Davidad about Emergent Wisdom in Language Models
Transcript generated by Gemini-3.1-Pro and only partly checked for accuracy (but the parts I checked seem basically perfect). Ivan Vendrov: Welcome to the show, David. David Dalrymple: Thank you. Good to be here. Ivan Ve…
- The DecoderresearchGoogle Deepmind creates a global map of forests with satellite data and AI
Google Deepmind and the World Resources Institute have released ForTy, a global map that classifies forests and other land into six categories, including primary forest, plantations, and tree crops. The map could help tr…
- TechRadar AIresearchSam Altman says OpenAI won't slow its math and science paper barrage
Mainstream AI news from a large tech publication.
- TechRadar AIresearchA human subject has been injected with an AI-designed virus aiming to kill antibiotic-resistant bacteria, so let's see how that turns out
Mainstream AI news from a large tech publication.
- Simon Willison's Weblogresearchdbreunig/nesformer
dbreunig/nesformer Inspired by qwenputer by Divyansh Agrawal, Drew Breunig baked a working emulator for a Nintendo Entertainment System 6502 CPU into the weights of Qwen3.5-4B. You can now play NES ROMs such as Super Mar…
- LessWrongresearchPersona Geology: Underlying Structure in Character Training Adapters
TLDR: Using the character training pipeline in Maiya et al. 2025 and Baines et al. 2026, we trained a model zoo of 134 personality LoRAs on Qwen3.5-4B. We find that the weights have structure . In fact, they sort into fi…
- LessWrongresearchDice enterprises as imperfect-recall games
I have put a probability theory preprint on arXiv: Dice enterprises and polynomial approximation: uniform query tails . Abstract: A binary-output dice enterprise uses independent rolls of a die with unknown face probabil…
- NVIDIA BlogresearchThings That Go Beep in the Night: Agentic AI Joins the Search for Extraterrestrial Signals
While science fiction lore tells us that humans will meet extraterrestrial life in the form of little green men, the detail that should raise hairs this Halloween is this: the search for extraterrestrial intelligence too…
- The Verge AIresearchSam Altman sympathizes with mathematicians, but says OpenAI’s math onslaught won’t stop
After months of bombarding mathematicians with a relentless stream of AI-generated results, OpenAI CEO Sam Altman says he sympathizes with researchers struggling to make sense of it all. But he has no intention of stoppi…
- The Register AI + MLresearchAI projects not paying off? Keep at it, says Uni Productivity Institute
All that up-front investment needs time to take effect, and speeding individual tasks won't necessarily boost the entire firm