Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
Google Research introduced TimesFM-3, a 330-million-parameter time series foundation model that performs multivariate forecasting in a single forward pass, pre-trained on over 1 trillion time points. The model, released …
An AI system developed by researchers at Imperial College London can detect heart failure and heart valve disease from a routine electrocardiogram in under two seconds, identifying up to 81% of heart failure cases and 90…
Doctors have developed a 'superhuman' AI tool that can spot heart disease in less than two seconds, trained on millions of routine electrocardiograms (ECGs) to extract more information than the human eye can see. The tec…
OpenShot 4.0, the latest version of the open-source video editor, has been released, introducing screen and webcam recording, color correction tools, and locally run machine learning models for subject isolation. The upd…
Priyank Jain argues that telecom operators approve machine learning models for deployment based on a single aggregate metric such as accuracy or AUC, which hides five failure modes that cause production failures. These i…
A developer's guide to building high-precision AI question-and-answer systems argues that basic vector search alone fails and recommends a multi-stage 'Retrieve and Re-rank' pipeline using a Cross-Encoder model to improv…
Google's Android 17 QPR2 Beta 4 adds a dedicated 'Data logging' menu for Private Compute Core within Privacy Dashboard, letting users monitor and manage outbound data from the secure sandbox that powers machine learning …
Decispher launched a persistent memory layer for coding agents that cuts token usage by 38×, according to its LongMemEval benchmarks, by organizing context from GitHub, documentation, and engineering platforms into reusa…
Cursor launched Router on July 22, 2026, a request-level classifier trained on over 600,000 live coding sessions that automatically selects the AI model for each prompt, cutting AI coding costs by up to 68%. In productio…
Google's LiteRT.js, a browser runtime for on-device AI inference using WebAssembly and WebGPU, enables developers to run standard .tflite models directly in the browser, as demonstrated in a tutorial for building an OCR …
A new paper from researchers including Multiverse Computing introduces Quantization-Aware Healing (QAH), a method that distills a full-precision model into a 4-bit compressed version, resulting in a model that outperform…
Towards Data Science published a framework for building retrieval-augmented generation (RAG) pipelines that introduces complexity only in response to observed failure modes, progressing from lexical and hybrid search to …
Amazon AGI researchers found that reinforcement learning fine-tuning with Group Relative Policy Optimization (GRPO) improves long-context model performance by up to 20 points on HotpotQA benchmarks and by 9 points over R…
A developer created a two-phase memory probe to evaluate AI code reviewers that retain context across pull requests, exposing a failure mode where bots trust stale memory over current repository decisions. The probe incl…
A developer is seeking a pipeline to convert STEM PDF textbooks into searchable embeddings, with stages for OCR-to-Markdown conversion, figure extraction, JSON schema structuring, and vector embeddings for Firestore Vect…
A tutorial demonstrates building a unified scikit-learn pipeline that combines text embeddings from Hugging Face's sentence-transformers with tabular features for classification, using the SMS Spam Collection dataset aug…
A century-old statistical method, Statistical Process Control (SPC), outperforms or matches modern deep learning models on the TSB-AD-M benchmark, exposing fundamental flaws in the benchmark's triviality, according to te…
The WUIC Assistant, an agentic VS Code plugin built on the WUIC framework, improved its local coding agent by applying Direct Preference Optimization (DPO) to a 30B-parameter model served via Ollama. The team fixed two e…
A developer argues that AI agents often fail not due to poor prompts but due to a lack of real tools, and demonstrates how equipping an agent with simple functions like running tests can dramatically improve performance.…
Jie Wang of Southwest University of Political Science and Law is seeking endorsement from arXiv cs.LG moderators for a paper titled 'Coupled Scaling: A Representational Accessibility Framework for Neural Scaling Laws,' w…