AI news archive.
10,660 articles filed under Research, newest first, from the outlets listed on the sources page.
- Apple Machine Learning ResearchresearchPlan or Retrieve First? A Scale Study of Tool Discovery Architectures
Tool-augmented LLM agents discover relevant APIs from large registries before planning execution, yet finite context windows prohibit injecting all tool definitions into a single prompt. The dominant approach, which we t…
- LessWrongresearchI trained GPT-2 with full-batch gradient descent so you don't have to
Note: The results of this are not as clean as they would naively appear, since a lot of training parameters are tuned for non-full-batch gradient descent. See discussion in the comments. I've left this up in case it's st…
- Latent SpaceresearchLab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio
How Leash Bio tamed DELs, the assay that traumatizes everyone, and used the data to catch cheating models. And why that same data is perfect for AI agents learning to think like chemists.
- Latent SpaceresearchLab-in-the-Loop at Compute Speed — Ian Quigley, Leash Bio
How Leash Bio tamed DELs, the assay that traumatizes everyone, and used the data to catch cheating models. And why that same data is perfect for AI agents learning to think like chemists.
- The DecoderresearchGoogle says self-improving AI isn't anywhere close
Google engineer Haydn Belfield says there's no measurable acceleration of AI research driven by AI at Deepmind. AI now writes more than 75 percent of production code, but 88 percent of changes still need human oversight.…
- WIRED AIresearchAI Came for Math. Now It’s Chasing the Rest of Science
AI may soon be the default way to solve theoretical problems in economics, high-energy physics, and other fields, MIT researchers tell WIRED.
- Ars Technica AIresearchResearchers use AI to design proteins that cut up other proteins
AI design tools make proteins that break bonds that are stable for hundreds of years.
- The Register AI + MLresearchResearchers cig-nal their faith in AI by taking up smoking
What are they puffing? Tobacco, apparently. Cancer researcher tells El Reg it's a bad bet
- Google ResearchresearchResearch@ 2026: Novel capabilities and intelligence
We’re sharing our latest milestones and some of the advanced AI capabilities and intelligence that we’re putting in the hands of experts.
- Hugging Face BlogresearchHiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost
Open-source model releases, tooling, and community updates.
- The Verge AIresearchMicrosoft has a new machine learning tool to improve weather forecasts
Old-school physics-based weather forecasting and even new AI models typically become less reliable beyond two weeks, a problem Microsoft claims to have solved using machine learning. Researchers with the company say they…
- LessWrongresearchPPA 11: The Basic Model
[This is the eleventh piece in my series on the Parametric Process Assumption (PPA), a new model of anthropics. You should start here if you are just seeing this for the first time. The previous piece is " Deriving the P…
- LessWrongresearchPPA 10: Deriving the PPA
[This is the tenth piece in my series on the Parametric Process Assumption (PPA), a new model of anthropics. You should start here if you are just seeing this for the first time. The previous piece is " Putting it All To…
- LessWrongresearchPPA 7: Temporal Consistency and Texas Sharpshooting
[This is the seventh piece in my series on the Parametric Process Assumption (PPA), a new model of anthropics. You should start here if you are just seeing this for the first time. The previous piece is " Indexical Model…
- LessWrongresearchPPA 6: Indexical Models Don't Have Diversity Biases
[This is the sixth piece in my series on the Parametric Process Assumption (PPA), a new model of anthropics. You should start here if you are just seeing this for the first time. The previous piece is " Improper Priors a…
- LessWrongresearchThe PPA: A New Approach to Anthropics
[This is the first piece in a series on the Parametric Parameter Assumption (PPA), a new model of anthropics. This particular piece is self-contained and for most people will contain all they need to know. TLDR: With the…
- LessWrongresearchA story on one possible future of AI-driven research
A quickly written hypothetical year-by-year of what happens to research as autonomous AI researchers become more capable. This is primarily meant to start conversation, I currently have very low confidence about the futu…
- AI Stack ExchangeresearchHow was $m ≥ 8log(\frac{7|H|}{6})$ derived here?
Section 5.1 of Understanding Machine Learning: From Theory to Algorithms 2014 by Shai Shalev-Shwartz and Shai Ben-David states: With probability of at least $\frac{1}{7}$ over the choice of $S ∼ D^m$ we have that $L_D(A(…
- WIRED AIresearchMen Can’t Stop Faking How Much They Know About AI
More than half of the men surveyed in a new Harvard study claimed to recognize AI leaders who were, in fact, totally fictional.
- Ai2 BlogresearchHiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost
HiRO-ACEv2 captures extreme rain, wind, and pressure events at 3 km resolution for about 1/300th the cost of X-SHiELD, making detailed global climate simulations more accessible to researchers.
- arXiv cs.AIresearchArtificiality's First Exam: An LLM-Referenced C-test for Supraintelligence
arXiv:2610.14005v1 Announce Type: new Abstract: Under the accelerated pace of AI development, any benchmark whose difficulty is calibrated with humans or current AI systems becomes circular, or saturates soon. How can we…
- arXiv cs.AIresearchCorrection Training for Diffusion Models
arXiv:2610.14136v2 Announce Type: replace Abstract: Iterative sampling in diffusion models can lead to error accumulation due to exposure bias. The denoiser is trained on real images interpolated with noise, but at infer…
- arXiv cs.AIresearchVETNET: Assembling Benchmark Datasets for AI-Driven Veterinary Innovation
arXiv:2610.14149v1 Announce Type: new Abstract: VETNET represents an ecosystem of live benchmark datasets and evaluation frameworks designed to accelerate AI-driven innovation in animal science and veterinary medicine th…
- arXiv cs.AIresearchCross-Modal Latent Influence: Patch-Level Training Sample Attribution for Vision-Language Models
arXiv:2610.14190v1 Announce Type: new Abstract: Explaining a vision--language model (VLM) prediction requires tracing training-example influence to the multimodal evidence that mediates it, yet existing attribution metho…
- arXiv cs.AIresearchEcho: Error-Compensated 4-Bit Tensor-Parallel Communication for LLM Training and Inference
arXiv:2610.14277v1 Announce Type: new Abstract: Tensor parallelism ends every layer with an all-reduce that sits between two dependent matrix multiplications, so its cost cannot be hidden and grows as compute outpaces in…
- arXiv cs.AIresearchControllable Diverse Sampling through Pullback Directions
arXiv:2610.14294v1 Announce Type: new Abstract: Conditional diffusion models can generate high-quality images that align closely with a given prompt. However, samples generated from different random seeds may remain visu…
- arXiv cs.AIresearchConicFlow: constrained Inference-time optimal control for flow matching models
arXiv:2610.14307v1 Announce Type: new Abstract: We propose ConicFlow, which frames constrained sampling from a flow matching model at inference time as an optimal control problem that balances deviation from the pretrain…
- arXiv cs.AIresearchActionable Meta-Requirements for AI Robustness Auditing
arXiv:2610.14356v1 Announce Type: new Abstract: AI-enabled systems must remain dependable when the conditions under which they operate change (e.g., due to input perturbations, distribution shifts, or changes in external…
- arXiv cs.AIresearchWhen Pleasing the User Compromises Care: AI Sycophancy in Caregiver-Mediated Healthcare
arXiv:2610.14360v1 Announce Type: new Abstract: Large language model (LLM) alignment often assumes that the user interacting with a model is also the primary beneficiary of its behavior. Yet in many consequential setting…
- arXiv cs.AIresearchMeasuring the Inference Energy Efficiency of Open-Weight LLMs across Benchmarks, Reasoning Modes, and Hardware
arXiv:2610.14389v1 Announce Type: new Abstract: Benchmark scores tell us what an LLM can do, but not how much energy is required to deliver that capability. We directly measure this trade-off for 86 open-weight model che…
- arXiv cs.AIresearchSAT-RL: Training Language Models for Logical Reasoning with Symbolic Solvers
arXiv:2610.14402v1 Announce Type: new Abstract: Large language models (LLMs) have achieved strong performance in mathematical and logical reasoning, yet they still suffer from the curse of complexity and struggle as prob…
- arXiv cs.AIresearchAlignPoison: Conditional Targeted Poisoning of Generative Recommenders via Decoding-Aligned Prefixes
arXiv:2610.14487v1 Announce Type: new Abstract: Generative recommendation models based on autoregressive decoding have been adopted as alternatives to classical ranking-based systems. While their modeling flexibility has…
- arXiv cs.AIresearchHiPPT: Physiology-Grounded Wearable PPG Representation Learning via Multi-Scope Text Alignment
arXiv:2610.14489v2 Announce Type: replace Abstract: Reusable wearable PPG representations should capture local pulse morphology and broader physiological regularities. Existing pretraining objectives, however, usually ap…
- arXiv cs.AIresearchReels-RSI: Program-Level Self-Improvement for Code-Rendered Explainer Videos
arXiv:2610.14495v1 Announce Type: new Abstract: Explainer videos -- a worked geometry problem, why the sky is blue, how a dynasty fell -- are usually one more target for pixel video generation. We argue they are a differ…
- arXiv cs.AIresearchMontageFlow: Physically Valid High Fidelity EEG Generation
arXiv:2610.14507v1 Announce Type: new Abstract: Generative modeling offers a promising solution to the scarcity of labeled clinical electroencephalography (EEG) data under strict privacy constraints. Although recent gene…
- arXiv cs.AIresearchScienceBuddy-PV: A Verifier-Governed Recursive Improvement Protocol for Panvascular Intervention Research
arXiv:2610.14523v1 Announce Type: new Abstract: ScienceBuddy-PV adapts the ScienceBuddy recursive-in-recursive design to panvascular intervention research through a structured evidence ledger, device-vessel ontology, abs…
- arXiv cs.AIresearchEnergy Ring: Stabilizing Feature Energy for Reliable Test-Time Discovery
arXiv:2610.14525v1 Announce Type: new Abstract: Test-Time Discovery (TTD) aims to recognize known classes and detect novel ones while the model is already deployed, operating directly on a continual unlabeled test-time s…
- arXiv cs.AIresearchFDE-Bench: Evaluating End-to-End Delivery from Underspecified Real-World Business Requests
arXiv:2610.14539v3 Announce Type: replace Abstract: Real-world business requests often leave key requirements unstated because frontline practitioners struggle to articulate rules they know from experience. Forward-deplo…
- arXiv cs.AIresearchAre Multimodal Benchmarks Really Useful? Item-Level Multimodal Benchmark Diagnosis via Structure-Response Co-Calibration
arXiv:2610.14542v2 Announce Type: replace Abstract: Multimodal benchmarks are central to evaluating whether models can reason over information from multiple modalities. However, existing literature has revealed that some…
- arXiv cs.AIresearchPsyMemBehav: Psychology-Informed and Memory-Grounded Two-Level Identity-to-Action Reasoning for Social-Media Behavior Simulation
arXiv:2610.14570v1 Announce Type: new Abstract: Large language models (LLMs) support social simulation, but faithfully reproducing real users requires grounding both structured platform decisions and language in sparse,…
- arXiv cs.AIresearchLearning after Execution with Action-Specific Credit
arXiv:2610.14573v1 Announce Type: new Abstract: Execution is often used as terminal feedback: it scores a generated program, but the resulting state is not used for another learning action. Learning after execution inste…
- arXiv cs.AIresearchIterate on an LLM, Serve on an SLM
arXiv:2610.14576v1 Announce Type: new Abstract: As frontier LLMs gain more capabilities, the cost to run them also goes up. At the same time, small language models (SLMs) also gain more capabilities but have largely been…
- arXiv cs.AIresearchDecoupling Knowledge Forgetting and Refusal Learning in MLLM Unlearning
arXiv:2610.14605v1 Announce Type: new Abstract: Current evaluations of multimodal unlearning typically focus on whether a model stops outputting target information. This surface-level criterion, however, masks two critic…
- arXiv cs.AIresearchAegisVox: Factorized Privacy and Clinical Preservation for Long-Form Speech
arXiv:2610.14643v1 Announce Type: new Abstract: Clinical speech reveals identity through both spoken identifiers and voice, yet its pauses, delivery, and speaker roles can also carry useful clinical evidence. Protecting…
- arXiv cs.AIresearchShaping Latent Mixture-of-Experts: From Expert Geometry to Layerwise Capacity Allocation
arXiv:2610.14644v1 Announce Type: new Abstract: Fine-grained mixture-of-experts (MoE) models commonly increase expert count by narrowing each expert, treating expert shape as a by-product of routing granularity. LatentMo…
- arXiv cs.AIresearchA Hybrid Kinematic GCN and Elastic Transformer for Efficient Indian Sign Language Recognition
arXiv:2610.14680v1 Announce Type: new Abstract: Sign Language Recognition (SLR) presents complex challenges due to variable signing speeds and spatial intricacies. Specifically, Indian Sign Language (ISL) requires precis…
- arXiv cs.AIresearchAcrossVAM2.0: Conflict-Aware Point Supervision for Parameter-Efficient Video Diffusion Adaptation
arXiv:2610.14693v1 Announce Type: new Abstract: Explicit point correspondence is attractive for adapting video generators: it describes how visible surfaces move, but can overwrite the appearance prior learned by a large…
- arXiv cs.AIresearchJoint Discrete and SE(3) Diffusion for Protein Interface Design
arXiv:2610.14699v1 Announce Type: new Abstract: The design of high-affinity, target-specific protein binders is central to the development of modern therapeutics. However, existing protein-generation methods remain limit…
- arXiv cs.AIresearchHow to Build a Benchmark That Actually Measures Taste
arXiv:2610.14707v2 Announce Type: replace Abstract: Expert-aligned "taste" benchmarks should reward craft judgment, not the ability to game the test. We introduce TasteBench, a pairwise good-vs-bad benchmark across visua…
- arXiv cs.AIresearchRelational Semantic Grounding for Test-Time Discovery
arXiv:2610.14721v1 Announce Type: new Abstract: Test-Time Discovery (TTD) aims to recognize known classes and detect novel ones from a continuous unlabeled test stream after deployment. Existing TTD methods mainly rely o…