AI news archive.
10,659 articles filed under Research, newest first, from the outlets listed on the sources page.
- arXiv cs.AIresearchPatients-like-me: A Variational LM--GNN Framework for Explainable Clinical Prediction
arXiv:2608.04193v2 Announce Type: replace-cross Abstract: Language models (LMs) offer strong textual representations for electronic health records (EHRs), but they encode patient sequences in isolation and provide limite…
- arXiv cs.AIresearchNeed or Reputation? Region, National Income, and Academic Status in LLM Gatekeeping Decisions
arXiv:2608.05178v2 Announce Type: replace-cross Abstract: Equitable access to scientific knowledge often depends on informal gatekeeping decisions, particularly when resources such as paywalled articles, nonpublic datase…
- arXiv cs.AIresearchFactorized Hypothesis Search for Evidence-to-Taxonomy Retrieval
arXiv:2608.06614v2 Announce Type: replace-cross Abstract: Large-taxonomy retrieval often assumes that the input already expresses the target concept. In many settings, however, the input is indirect evidence, such as a t…
- arXiv cs.AIresearchLifelongCrossNav: Persistent 3D Semantic Memory for Cross-Floor Multi-Object Navigation
arXiv:2608.07079v2 Announce Type: replace-cross Abstract: Object-goal navigation has made substantial progress in semantic perception and exploration, yet persistent memory for multi-object navigation and cross-floor nav…
- arXiv cs.AIresearchMeasuring Concept Content in Text from LLM Activations: ESG Evidence from Concept Vectors and Linear Probes
arXiv:2608.07208v2 Announce Type: replace-cross Abstract: Recent work shows that Large Language Models (LLMs) encode more information in their activations than they show in their responses. We ask whether the information…
- arXiv cs.AIresearchLost in Context Compaction: When LLM Systems Drop Side Constraints
arXiv:2608.11242v2 Announce Type: replace-cross Abstract: When the context window is under pressure, LLM systems compact prior context to continue ongoing tasks. We study Side Constraints (SCs), user-issued instructions…
- arXiv cs.AIresearchA, B, C, or Stable? Accuracy and Order Sensitivity Diverge in LLM Multiple-Choice Evaluation
arXiv:2608.11947v2 Announce Type: replace-cross Abstract: Multiple-choice question (MCQ) benchmarks are widely used to evaluate large language models, yet a single reported accuracy number is not enough to distinguish a…
- arXiv cs.AIresearchP2Skill: Privacy Preserving Skill Distillation for Cloud-Local LLM Inference Systems
arXiv:2608.14094v2 Announce Type: replace-cross Abstract: Cloud-local LLM inference can use the reasoning capability of large cloud models, but cloud-bound requests must exclude personally identifiable information (PII).…
- arXiv cs.AIresearchPlausible but Not Valid: A Psychometric Audit of LLMs as Synthetic Survey Respondents
arXiv:2608.14606v2 Announce Type: replace-cross Abstract: Large language models (LLMs) are increasingly used as synthetic survey respondents, but existing evaluations mostly ask whether individual answers look plausible.…
- arXiv cs.AIresearchSIDScope: A Diagnostic Resource for Semantic-ID Interfaces in Generative Recommendation
arXiv:2608.18779v2 Announce Type: replace-cross Abstract: Semantic ID tokenizers assign code sequences to catalog items, which generative recommenders predict and resolve back to items. Reusing a released mapping require…
- arXiv cs.AIresearchDualOPSD: Closing the Feedback Loop in On-Policy Self-Distillation
arXiv:2608.26019v2 Announce Type: replace-cross Abstract: On-policy self-distillation (OPSD) combines student generated trajectories with dense supervision from a privileged teacher of the same model. Its pointwise clipp…
- arXiv cs.AIresearchDelegating Before Learning: Where Generative AI Sits in Students' Professional Communication
arXiv:2608.28837v2 Announce Type: replace-cross Abstract: We conducted an interview study with twelve students on their use of generative AI in academic communication. Students delegated professional messages to AI most…
- arXiv cs.AIresearchLatentPress: Context Compression Beyond Text and Vision
arXiv:2609.01507v3 Announce Type: replace-cross Abstract: Compressed context is usually carried as human-readable text or as rendered images that must be decoded, even when its consumer is a language model. We introduce…
- arXiv cs.AIresearchTC-Next: Zero-Shot Multimodal Cyclone Forecasting
arXiv:2609.02085v2 Announce Type: replace-cross Abstract: We present TropicalCycloneNext (TC-Next), a multimodal deep learning model that forecasts tropical cyclone track and intensity at $6-24$h leads by leveraging a fo…
- arXiv cs.AIresearchFrom Portraits to Pathology: Probing Diagnostic Overreach in Vision-Language Models
arXiv:2609.05540v2 Announce Type: replace-cross Abstract: Many medical conditions require clinical assessment of symptoms, history, and functioning rather than judgments from appearance. We study whether vision-language…
- arXiv cs.AIresearchSeRV: Semantic-Aligned Residual Vector Quantization for American Sign Language Generation
arXiv:2609.05742v2 Announce Type: replace-cross Abstract: American Sign Language (ASL) generation remains challenging due to limited paired text-ASL motion data and the difficulty of learning motion representations both…
- arXiv cs.AIresearchOpen Tabular Insight Extraction: Where Do We Stand, and Where Should We Go?
arXiv:2609.07629v2 Announce Type: replace-cross Abstract: Democratizing access to the knowledge held in large corpora of tables such as data lakes is emerging as a central research challenge. Research in this space is ad…
- arXiv cs.AIresearchRoboCousin: Build Your Own Simulation Playground for Robust Bimanual Robotic Manipulation
arXiv:2609.08339v2 Announce Type: replace-cross Abstract: Bimanual manipulation policies require large and diverse training datasets, yet collecting demonstrations on physical robots is expensive and difficult to scale.…
- arXiv cs.AIresearchSPINE: Measuring LLM Sycophancy under Sustained Multi-Turn Pressure
arXiv:2609.09090v2 Announce Type: replace-cross Abstract: Large language models (LLMs) may abandon correct positions when users push back, exhibiting a failure mode known as sycophancy. Existing evaluations typically use…
- arXiv cs.AIresearchDivide, Consult, Conquer: Capability Laundering Through Aligned LLMs
arXiv:2609.15383v2 Announce Type: replace-cross Abstract: Language model safety is typically evaluated one interaction at a time. We show that a weaker, unaligned model can split a harmful task into benign-looking subpro…
- Apple Machine Learning ResearchresearchThe Design Space of Tri-Modal Masked Diffusion Models
Discrete diffusion models have emerged as strong alternatives to autoregressive language models, with recent work initializing and finetuning a base unimodal model for bi-modal generation. Diverging from previous approac…
- Apple Machine Learning ResearchresearchCode-Switching Detection via Language-Selective Dimensions in Speech Foundation Models
Code-switching (CS), the alternation between languages within an utterance, is common in multilingual speech but underrepresented in pretraining corpora. Increasing CS coverage requires finding relevant utterances in lar…
- TechCrunch AIresearchScience Corporation implants its vision-restoring chip in its first Stargardt disease patient
Max Hodak, the former president and co-founder of Neuralink, detailed onstage at TechCrunch Disrupt how his company's work with PRIMA is part of its path toward a novel brain-computer interface (BCI).
- Latent SpaceresearchDyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Our first robotics pod with Dyna’s Chief Scientist on their big bet on World Action Models & the Next Frontier of Physical AI
- Latent SpaceresearchDyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics
Our first robotics pod with Dyna’s Chief Scientist on their big bet on World Action Models & the Next Frontier of Physical AI
- Ars Technica AIresearchThey war-gamed dimming the Sun. It went even worse than you think.
Solar geoengineering simulation shows what happens when geopolitics takes over.
- The DecoderresearchWestern mathematicians fear AI makes them dumber while China builds an entire alliance for it
While 25 Fields Medal winners in the West warn against AI-generated proofs and parts of the community call for a boycott, China is going on the offensive. Fields Medal winner Shing-Tung Yau has launched an alliance of ei…
- LessWrongresearchPersonas in a Pre-Internet Language Model
David Africa & Nick Levine TLDR We ran several persona experiments on Talkie, a 13B model trained only on English text from before 1931. We ran them alongside its modern twin (same architecture, trained on regular data a…
- LessWrongresearchWhen are safe Pareto improvements individually rational?
(Soft prerequisites: A high-level model of AI bargaining and Taboo “equilibrium”: Less confused frames for research on AI bargaining .) Section II.1 of CLR’s safe Pareto improvements (SPI) agenda briefly reviews the chal…
- LessWrongresearchA Pairing Strategy That Almost Works
Continues from my previous short form. https://yoonwoojeong.github.io/SYZYGY Here is a small puzzle. Two players each control one piece on a square grid. On your turn you either take a step, or you knock out one of the s…
- Apple Machine Learning ResearchresearchMix, Don't Tune: Bilingual Pre-Training Outperforms Hyperparameter Search in Data-Constrained Settings
For most languages of the world, language model pre-training operates in a data-constrained regime where models must repeat their training data many times, degrading generalization. Two remedies exist: aggressive hyperpa…
- LessWrongresearchGood Evals Are Hard: An ImpossibleBench Audit
(This post is a summary. See GitHub Repo for details .) Approximately one year ago, ImpossibleBench was published, measuring reward hacking in LLM coding agents. I ran the LiveCodeBench (LCB) half of ImpossibleBench on c…
- TechRadar AIresearchHas OpenAI started the 'mathpocalypse'? Experts say AI's solving of iconic math problems may cause huge issues in the long run
Mainstream AI news from a large tech publication.
- LessWrongresearchManifold Steering from a Natural State is Effective and Generalizable, but Relies on Model Accuracy
tl;dr I continue work to apply Goodfire lab's manifold steering methods to the Gemma-2-2b model's natural state, both in raw residual space and in the space produced by Anthropic's pre-trained CLT features. I find that t…
- The DecoderresearchSakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots
Sakana AI's MASS method lets one language model improve itself by solving tasks, judging results, and refining workflows, reaching a win rate of nearly 70 percent against the base model after two rounds. But when the mod…
- Google ResearchresearchApplying technology to understand and treat disease
In collaboration with scientists and research partners, Google is helping to make disease more detectable, treatable, and preventable — with the goal of improving health…
- LessWrongresearchThe Expensive-Loop Problem
LLMs are amazing at tasks where they have tight and legible feedback loops. Domains where experiments are quick, deterministic and cheap. I will call these domains "cheap-loop" domains. Presumably, this is mostly a resul…
- Simon Willison's WeblogresearchProfiling web-edit with Python 3.15: flame graphs and source heatmaps
Research: Profiling web-edit with Python 3.15: flame graphs and source heatmaps I put the new Python 3.15 profiler through its paces by having GPT-6 Astra Extra High in Codex run it against my new web-edit app : I need a…
- The DecoderresearchClearing chat history helps AI agents find better solutions
Microsoft researchers found that autonomous AI agents get trapped in what they call "idea basins" because of their growing chat context, local solution regions they can't escape. Their "Fork-and-Flush" method clears the…
- NVIDIA Developer BlogresearchReplaySSM: Rethinking State Storage for Faster Hybrid-Model Decoding
Hybrid models interleave state-space model (SSM) layers with a few attention layers. NVIDIA Nemotron 3, Qwen3.5 through Qwen3.8, Kimi-K3, and GLM-5.3-Flash are...
- Microsoft ResearchresearchSimulating student learning for better AI tutoring
StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners. It models how individuals may engage with tutors, giving feedback to improve tutor performance across s…
- LessWrongresearchQuantity vs quality in pottery is a lie
Epistemic status: I traced one famous anecdote to its printed source and to both coauthors' own accounts of where it came from. Confident it was never a study. No claim either way about whether quantity beats quality in…
- LessWrongresearchBase Models Can Detect Hidden Intentions that Prompted Classifiers Miss
TL;DR The Persona Selection Model (PSM) posits that post-training selects a persona the model learned during pre-training. Subliminal Learning ( Cloud et al.) showed that models can impart covert signals (like "love of o…
- LessWrongresearchMore research needed?
In an earlier post I showed that when AI systems can perform a task, they do it much more cheaply than a human worker. If, as I think is likely, this remains true as we develop AGI, today's stock of compute could run the…
- Import AIresearchImport AI 476: Creativity and RSI; worse-than-China open weight models; AI goes to space
Will we give robots the moon?
- TechRadar AIresearchWhen one robot learns, every robot knows if they can communicate
Mainstream AI news from a large tech publication.
- IEEE Spectrum AIresearchLearning to Play Like a Pro: Adaptive Motion Planning and Tracking for Humanoid Robots
This White Paper provides roboticists and control engineers with a comprehensive overview of how motion planning and motion tracking can be decoupled to preserve professional athletic style in humanoid robots, and why ad…
- The DecoderresearchGoogle Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed
Google Deepmind's AlphaProof Nexus solved dozens of open math problems, including nine Erdős problems two of which were unsolved for 56 years, per a study in Science. The publication comes as OpenAI faces a boycott for d…
- LessWrongresearchManifolds for Five-Year Olds
Ambitious mechanistic interpretability research is full of jargon, but the underlying intuitions are so simple that a five-year-old (with a firm grasp of linear algebra) could understand them. In one recent paper from Go…
- arXiv cs.AIresearchFaithfulness Without Anchors: A Position Paper on the Structural Misalignments of Explainable AI Faithfulness Metrics for Feature Importance
arXiv:2610.12493v1 Announce Type: new Abstract: Explainable Artificial Intelligence (XAI) tools have often been presented as methods to increase trust in machine learning models by making aspects of their algorithms more…