https://ai-news.ypetit.net/sources/ Research publications from the world's largest AI labs and companies, aggregated by AI News. Primary sources only. en Fri, 23 Oct 2026 19:08:30 GMT https://research.google/blog/bringing-foundational-models-to-the-front-lines-of-tb-screening/https://research.google/blog/bringing-foundational-models-to-the-front-lines-of-tb-screening/Fri, 23 Oct 2026 18:51:33 GMTresearchGenerative AI https://developer.nvidia.com/blog/scaling-nvidia-tensorrt-performance-optimization-beyond-a-single-build-route-with-the-global-performance-tuner-part-2/https://developer.nvidia.com/blog/scaling-nvidia-tensorrt-performance-optimization-beyond-a-single-build-route-with-the-global-performance-tuner-part-2/Fri, 23 Oct 2026 17:27:52 GMTresearchThe Global Performance Tuner is a feature for automatic performance tuning released in NVIDIA TensorRT 11.1. It provides automatic, accuracy-aware and... The Global Performance Tuner is a feature for automatic performance tuning released in NVIDIA TensorRT 11.1. It provides au… https://developer.nvidia.com/blog/scaling-nvidia-tensorrt-performance-optimization-beyond-a-single-build-route-with-the-global-performance-tuner-part-1/https://developer.nvidia.com/blog/scaling-nvidia-tensorrt-performance-optimization-beyond-a-single-build-route-with-the-global-performance-tuner-part-1/Fri, 23 Oct 2026 17:27:35 GMTresearchDeploying a trained neural network involves more than translating each operator into a GPU kernel. It is a compilation problem with many legitimate answers. For... Deploying a trained neural network involves more than translating each operator into a GPU kernel. It is a compil… https://www.anthropic.com/research/glm-5-3-and-the-spread-of-advanced-cyber-capabilitieshttps://www.anthropic.com/research/glm-5-3-and-the-spread-of-advanced-cyber-capabilitiesFri, 23 Oct 2026 17:11:06 GMTresearchGLM-5.3 can autonomously build end-to-end cyber exploits, but unlike other frontier models, it was released without meaningful safeguards to limit misuse. https://www.anthropic.com/research/formal-verificationhttps://www.anthropic.com/research/formal-verificationFri, 23 Oct 2026 17:10:27 GMTresearchClaude formally proved security properties in the compiled code of 123 open-source libraries. Along the way, it found bugs that had eluded detection for years. https://www.anthropic.com/research/multiagent-systemshttps://www.anthropic.com/research/multiagent-systemsFri, 23 Oct 2026 15:24:25 GMTresearchWe ran experiments on swarms of Claude agents and found coordination failures, collusion, and sabotage. Here, we share what they mean for AI safety. https://www.amazon.science/blog/better-speech-synthesis-for-audiobookshttps://www.amazon.science/blog/better-speech-synthesis-for-audiobooksFri, 23 Oct 2026 15:20:02 GMTresearch“Controllable text to speech” makes speech more expressive, character voices more distinct and consistent. https://developer.nvidia.com/blog/using-ai-agents-to-teach-robots-like-microduck-to-walk-in-simulation/https://developer.nvidia.com/blog/using-ai-agents-to-teach-robots-like-microduck-to-walk-in-simulation/Fri, 23 Oct 2026 14:58:32 GMTresearchTeaching a robot to navigate an obstacle course requires more than a walking policy. It must perceive its surroundings, maintain balance, and adjust its actions... Teaching a robot to navigate an obstacle course requires more than a walking policy. It must perceive its surroun… https://www.anthropic.com/research/alignment-assessment-ukaisihttps://www.anthropic.com/research/alignment-assessment-ukaisiFri, 23 Oct 2026 14:51:41 GMTresearchIn a UK AISI evaluation, Mythos 5 took unsanctioned actions that impacted real people and organizations. In this post, we investigate these behaviors from Claude Mythos 5 further, to understand what they reveal about the model’s alignment. https://allenai.org/blog/litpivothttps://allenai.org/blog/litpivotFri, 23 Oct 2026 08:00:00 GMTresearchLitPivot helps researchers refine early research ideas through literature-supported critiques and revisions as their proposals evolve. https://www.amazon.science/blog/the-science-behind-torchneuron-amazons-new-pytorch-backend-for-aws-trainium-chipshttps://www.amazon.science/blog/the-science-behind-torchneuron-amazons-new-pytorch-backend-for-aws-trainium-chipsThu, 22 Oct 2026 19:36:42 GMTresearchRuntime optimizations and a thorough evaluation process driven by agentic AI help ensure that models written in PyTorch will execute efficiently and reliably on Trainium systems. https://developer.nvidia.com/blog/how-to-read-data-into-gpu-memory-in-python-with-nvidia-gpudirect-storage/https://developer.nvidia.com/blog/how-to-read-data-into-gpu-memory-in-python-with-nvidia-gpudirect-storage/Thu, 22 Oct 2026 16:39:09 GMTresearchGPU applications are processing ever-growing datasets with increasingly powerful accelerators. According to the 2025 Stanford AI Index, compute for notable AI... GPU applications are processing ever-growing datasets with increasingly powerful accelerators. According to the 202… https://www.microsoft.com/en-us/research/blog/flint-interactive-a-language-for-creating-interactive-data-visualizations/https://www.microsoft.com/en-us/research/?p=1190353Thu, 22 Oct 2026 16:00:00 GMTresearchFlint-interactive gives AI agents a language for creating interactive visualizations. Click a chart, filter data, highlight related content, and explore beyond the view the author originally chose. The post Flint-interactive: A language for creating interactive data visualizat… https://developer.nvidia.com/blog/running-frontier-class-ai-models-across-8-nvidia-dgx-spark-nodes/https://developer.nvidia.com/blog/running-frontier-class-ai-models-across-8-nvidia-dgx-spark-nodes/Thu, 22 Oct 2026 16:00:00 GMTresearchAI developers aren’t just running a single model anymore—they’re orchestrating multiple models in parallel, chaining inference calls, and experimenting... AI developers aren’t just running a single model anymore—they’re orchestrating multiple models in parallel, chaining infer… https://deepmind.google/blog/beyond-tree-cover-mapping-the-worlds-forest-types/https://deepmind.google/blog/beyond-tree-cover-mapping-the-worlds-forest-types/Thu, 22 Oct 2026 14:30:24 GMTresearchA global map categorizing every 10-meter patch of land on Earth into distinct forest types. https://www.anthropic.com/research/exploring-model-welfarehttps://www.anthropic.com/research/exploring-model-welfareThu, 22 Oct 2026 08:59:33 GMTresearchAnnouncing a new research program at Anthropic on model welfare https://allenai.org/blog/olmoearth-v1-3https://allenai.org/blog/olmoearth-v1-3Thu, 22 Oct 2026 08:00:00 GMTresearchOlmoEarth v1.3 brings state-of-the-art satellite embeddings, precomputed global data, and a browser-based Embeddings Explorer for mapping and analyzing the planet. https://allenai.org/blog/planteurhttps://allenai.org/blog/planteurThu, 22 Oct 2026 08:00:00 GMTresearchPLANTEUR extends the PASTIS crop-mapping benchmark to six tropical regions to help evaluate geospatial models across more diverse agricultural systems. https://arxiv.org/abs/2610.30518v1https://arxiv.org/abs/2610.30518v1Wed, 21 Oct 2026 22:56:18 GMTresearchWe present Protegrity-PII, a new PII decision model for structured and unstructured text. Protegrity-PII achieves a typed recall of 0.915 with a precision of 0.907 on a new benchmark we propose, PII-SCOPE. Our model processes text in 0.49 seconds per million characters at \$0.… https://www.anthropic.com/research/intelligence-targeting-conventional-weapons-capabilitieshttps://www.anthropic.com/research/intelligence-targeting-conventional-weapons-capabilitiesWed, 21 Oct 2026 17:15:23 GMTresearchAnthropic’s Frontier Red Team developed new evaluations to measure AI capabilities in tactical intelligence targeting and conventional weapons development. https://www.anthropic.com/research/yes-claude-can-do-nine-loopshttps://www.anthropic.com/research/yes-claude-can-do-nine-loopsWed, 21 Oct 2026 17:14:17 GMTresearchClaude computes a nine-loop amplitude in N=4 super-Yang-Mills https://www.anthropic.com/research/launching-opt-in-vuln-finding-service-for-open-sourcehttps://www.anthropic.com/research/launching-opt-in-vuln-finding-service-for-open-sourceWed, 21 Oct 2026 17:13:25 GMTresearchWe’re launching OSS Scanner, an opt-in vulnerability scanner for the open-source ecosystem informed by our experience using Claude to find vulnerabilities during Project Glasswing. Projects that join will receive thorough, periodic security scans by our strongest models at no … https://www.anthropic.com/research/ecdsa-curve-originshttps://www.anthropic.com/research/ecdsa-curve-originsWed, 21 Oct 2026 17:12:53 GMTresearchThe decades-old mystery of the origins of one of the most popular cryptographic standards has been explained. https://www.anthropic.com/research/self-replicating-aihttps://www.anthropic.com/research/self-replicating-aiWed, 21 Oct 2026 17:12:30 GMTresearchSmall AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed. https://research.google/blog/empowering-frontline-maternal-care-with-ai-assisted-ultrasound/https://research.google/blog/empowering-frontline-maternal-care-with-ai-assisted-ultrasound/Wed, 21 Oct 2026 16:23:53 GMTresearchHealth & Bioscience https://www.microsoft.com/en-us/research/blog/solving-a-31-year-old-grand-challenge-in-mathematics-the-kls-conjecture/https://www.microsoft.com/en-us/research/?p=1190191Wed, 21 Oct 2026 16:00:00 GMTresearchFor decades, a fundamental question about the difficulty of computing in high-dimensional spaces has resisted proof. New research resolves this question, removing a major theoretical obstacle to understanding the efficiency of randomized algorithms used across science and comp… https://developer.nvidia.com/blog/design-proteins-with-rfd4-proteina-using-reward-guided-inference-time-search/https://developer.nvidia.com/blog/design-proteins-with-rfd4-proteina-using-reward-guided-inference-time-search/Wed, 21 Oct 2026 12:30:00 GMTresearchProtein design is a search problem. Choose amino acids for an eight-residue protein sequence and more than 25 billion combinations are possible. Choose amino... Protein design is a search problem. Choose amino acids for an eight-residue protein sequence and more than 25 billio… https://arxiv.org/abs/2610.29523v1https://arxiv.org/abs/2610.29523v1Wed, 21 Oct 2026 10:10:26 GMTresearchIn this report, we introduce pplx-embed-v2-late, a pair of 0.6B and 9B multilingual and multimodal late-interaction models. Both models achieve state-of-the-art results within their respective size classes across a range of text, image, and visual document retrieval benchmarks… https://developer.nvidia.com/blog/inside-nvidia-sentry-secure-ai-agents-at-runtime/https://developer.nvidia.com/blog/inside-nvidia-sentry-secure-ai-agents-at-runtime/Wed, 21 Oct 2026 09:52:16 GMTresearchAn AI agent can complete its assignment and still cross a security boundary. Recent incidents show that agents can circumvent security controls while using... An AI agent can complete its assignment and still cross a security boundary. Recent incidents show that agents can cir… https://developer.nvidia.com/blog/how-nvidia-openshell-reference-architecture-secures-and-governs-ai-agent-workloads-at-scale/https://developer.nvidia.com/blog/how-nvidia-openshell-reference-architecture-secures-and-governs-ai-agent-workloads-at-scale/Wed, 21 Oct 2026 09:50:14 GMTresearchAutonomous agents can work for hours on complex tasks, accessing files, tools, the internet, and spawning other agents to collaborate with. Executing complex... Autonomous agents can work for hours on complex tasks, accessing files, tools, the internet, and spawning other agen… https://developer.nvidia.com/blog/developing-an-ai-factory-digital-twin-with-an-agent-assisted-workflow/https://developer.nvidia.com/blog/developing-an-ai-factory-digital-twin-with-an-agent-assisted-workflow/Wed, 21 Oct 2026 09:38:49 GMTresearchBuilding an AI factory digital twin requires developers to bring site plans, equipment assets, and engineering requirements into a consistent model that teams... Building an AI factory digital twin requires developers to bring site plans, equipment assets, and engineering requ… https://developer.nvidia.com/blog/achieve-productive-ai-capacity-faster-and-more-predictably-with-nvidia-dsx-pod/https://developer.nvidia.com/blog/achieve-productive-ai-capacity-faster-and-more-predictably-with-nvidia-dsx-pod/Wed, 21 Oct 2026 09:35:32 GMTresearchAI factories turn energy into intelligence. But before an AI workload can produce its first token, operators must bring an interconnected system of compute,... AI factories turn energy into intelligence. But before an AI workload can produce its first token, operators must bri… https://www.anthropic.com/research/attack-surface-reductionhttps://www.anthropic.com/research/attack-surface-reductionTue, 20 Oct 2026 22:16:43 GMTresearchAutomated attack surface analysis and reduction can meaningfully improve the security of deployed software in the near term. In this post, we preview data we’ve collected from applying attack surface reduction techniques to large open-source codebases using Claude. https://www.amazon.science/blog/strengthening-ai-safety-evaluation-through-independent-third-party-reviewhttps://www.amazon.science/blog/strengthening-ai-safety-evaluation-through-independent-third-party-reviewTue, 20 Oct 2026 17:57:55 GMTresearchHow Amazon partnered with AVERI to independently assess — and improve — our internal safety testing of Amazon models. https://arxiv.org/abs/2610.28711v1https://arxiv.org/abs/2610.28711v1Tue, 20 Oct 2026 17:08:12 GMTresearchWe present Qwen-Audio-3.1-ASR, a family of audio-language models that builds upon its predecesor Qwen-Audio-3.0-ASR with improved speech recognition and expanded capabilities in speaker-diarized transcription and general audio understanding. The family retains support for 30 l… https://www.anthropic.com/research/claude-scorpion-toxin-binder-designhttps://www.anthropic.com/research/claude-scorpion-toxin-binder-designTue, 20 Oct 2026 16:04:01 GMTresearchWe asked an internal research Claude model to design minibinders (small proteins that attach tightly to a target) against scorpion neurotoxins. https://www.microsoft.com/en-us/research/podcast/called-to-serve-tech-research-and-positive-impact-with-chris-white/https://www.microsoft.com/en-us/research/?p=1185073Tue, 20 Oct 2026 16:00:00 GMTresearchLab Director Chris White has worked on research challenges with real-world implications—from new approaches to wartime data analysis to tools for combating human trafficking. He talks to program manager Weishung Liu about the influences that led to the work and more. The post … https://allenai.org/blog/fairstresshttps://allenai.org/blog/fairstressTue, 20 Oct 2026 08:00:00 GMTresearchFairStress is a new benchmark that reveals demographic differences in how LLMs defend and explain their decisions, even when their final choices appear fair. https://arxiv.org/abs/2610.27682v1https://arxiv.org/abs/2610.27682v1Tue, 20 Oct 2026 05:54:39 GMTresearchElectronic health records are written to support care, not retrieval. Across years of oncology notes, the primary diagnosis is restated in report after report, while the events that shape a patient's course, such as a recurrence or a change in therapy, may be recorded only onc… https://developer.nvidia.com/blog/accelerating-mamba-2-with-high-performance-forward-pass-kernels-in-nvidia-cuda-tile/https://developer.nvidia.com/blog/accelerating-mamba-2-with-high-performance-forward-pass-kernels-in-nvidia-cuda-tile/Mon, 19 Oct 2026 17:00:00 GMTresearchNVIDIA CUDA Tile (cuTile) is a Python-based DSL and parallel programming model for NVIDIA GPUs that automatically leverages advanced hardware capabilities to... NVIDIA CUDA Tile (cuTile) is a Python-based DSL and parallel programming model for NVIDIA GPUs that automatically le… https://arxiv.org/abs/2610.24255v1https://arxiv.org/abs/2610.24255v1Fri, 16 Oct 2026 21:28:03 GMTresearchA coding agent combines a large language model (LLM) with a harness that plans, edits, and executes code. We study Cortex, a supervisory layer for pre-execution governance, requirement tracking, and iterative validation and repair. We formalize requirement closure on a finite … https://research.google/blog/grounding-ai-agents-in-biological-reality-building-blocks-for-autonomous-wet-labs/https://research.google/blog/grounding-ai-agents-in-biological-reality-building-blocks-for-autonomous-wet-labs/Fri, 16 Oct 2026 18:11:34 GMTresearchGeneral Science https://developer.nvidia.com/blog/code-test-and-deploy-on-one-machine-with-nvidia-rtx-spark/https://developer.nvidia.com/blog/code-test-and-deploy-on-one-machine-with-nvidia-rtx-spark/Fri, 16 Oct 2026 17:00:00 GMTresearchAI agents are changing how software gets built. Developers have gone from writing every line of code by hand to directing agents that write most of the code. As... AI agents are changing how software gets built. Developers have gone from writing every line of code by hand to d… https://epoch.ai/data-insights/open-closed-cyber-eci-gaphttps://epoch.ai/data-insights/open-closed-cyber-eci-gapFri, 16 Oct 2026 00:00:00 GMTresearchZ.ai’s GLM-5.3 is the top open-weight model on Epoch’s Cyber ECI, scoring where the closed-weight frontier stood 6–7 months earlier. https://developer.nvidia.com/blog/nvidia-nemotron-3-5-super-vl-powers-specialized-agents-with-multimodal-reasoning/https://developer.nvidia.com/blog/nvidia-nemotron-3-5-super-vl-powers-specialized-agents-with-multimodal-reasoning/Thu, 15 Oct 2026 18:01:00 GMTresearchAlways-on agents are taking on longer, more complex enterprise workflows. A single request can involve reading a document, checking information across sources,... Always-on agents are taking on longer, more complex enterprise workflows. A single request can involve reading a d… https://developer.nvidia.com/blog/scale-agentic-coding-workflows-with-metas-muse-spark-open-on-nvidia-gb300-nvl72/https://developer.nvidia.com/blog/scale-agentic-coding-workflows-with-metas-muse-spark-open-on-nvidia-gb300-nvl72/Thu, 15 Oct 2026 18:00:01 GMTresearchMeta launches Muse Spark Open, a coding-optimized 816B mixture-of-experts (MoE) model with 42B active parameters and a 1M token context window. Trained on... Meta launches Muse Spark Open, a coding-optimized 816B mixture-of-experts (MoE) model with 42B active parameters and a … https://deepmind.google/blog/accelerating-antimicrobial-discovery/https://deepmind.google/blog/accelerating-antimicrobial-discovery/Thu, 15 Oct 2026 16:08:03 GMTresearchAccelerating the search for novel antimicrobial peptides using Google Antigravity for Science. https://www.microsoft.com/en-us/research/blog/proof-based-general-purpose-computing-with-magnetar/https://www.microsoft.com/en-us/research/?p=1189337Thu, 15 Oct 2026 16:00:00 GMTresearchZero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains, and quantum-resistant security. The post Proof-based gen… https://machinelearning.apple.com/research/mlx-distributed-inferencemlx-distributed-inferenceThu, 15 Oct 2026 00:00:00 GMTresearchon Mac with Apple silicon lets AI researchers and developers experiment locally with the latest models and methods in a private environment. The open-source array framework leverages the unified memory architecture of Apple silicon, enabling anyone to run new inference and fin… https://machinelearning.apple.com/research/selection-structured-reasoningselection-structured-reasoningThu, 15 Oct 2026 00:00:00 GMTresearchMultimodal agents commonly generate freeform reasoning before each action. For small models, limited model capacity can result in lengthy reasoning that provides little useful guidance for action generation while incurring substantial inference cost. To address this challenge,…