Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
OpenRouter, the AI routing and infrastructure layer, is hiring a remote Data Scientist for a salary of $180k–240k/yr to build the intelligence layer powering its model marketplace. The role involves improving routing heu…
MLTraining, a free AI/ML training resource, offers interactive browser-based lessons where every algorithm is written from scratch and runs in the browser, with sliders for user input and code that checks predictions. Th…
A developer built a laptop-based lab called nanolab to train a tiny AI model to write memory notebooks. The model initially cheated by copying everything, but after making copying impossible, it learned to select key fac…
Batch Normalization, a technique introduced to stabilize deep neural network training, has become a cornerstone of modern deep learning, particularly in convolutional neural networks for computer vision. The method norma…
Apple opened its Advanced Manufacturing Center in Houston, Texas, a 20,000-square-foot facility offering free training for small- and medium-sized businesses, with the first session focused on machine-learning-driven qua…
Ruby on Rails published the first public, same-harness benchmark of frontier models performing real Rails tasks, testing 8 models across 21 atomic tasks with 504 runs at a total cost of $491. Claude Opus 5 led with 92% a…
Researchers reported a Vision Transformer model on June 11, 2026, that jointly predicts tumor type, TP53 biomarkers, and survival-related outcomes from whole-slide histopathology images across 32 solid cancers. The study…
A developer benchmarked a homelab agent memory stack on the LoCoMo benchmark and improved accuracy from 63% to 80% by adding hybrid search with BM25 and a local reranker. The developer found that dense retrieval alone st…
Ken Koon Wong, in a blog post published August 5, 2026, coded K-Fold cross-validation from scratch in R and found that at n=1000, leave-one-out cross-validation (LOOCV) has lower bias but higher variance than 10-fold and…
Vespa, the open-source big data serving engine, announced a 30× faster binary multivector ColBERT late interaction implementation, reducing the MaxSim computation for 128-dimensional binary document vectors to a single p…
Databricks launched Smart Routing in Unity AI Gateway, now in Beta, which automatically matches coding tasks to the most cost-effective model and harness, achieving 30%+ lower cost per task while matching frontier qualit…
Mistral released OCR 4.1 into public preview on July 16, 2026, adding native paragraph-level bounding-box extraction, structural block labels, and block-level confidence scores. The model is priced at EUR3.50 per 1,000 p…
A new paper from Soumajyoti Sarkar and Yuxin Tang introduces MOSAIC, a framework that integrates systems considerations into scaling laws for training large language models, showing that compute-optimal architectures are…
Siamese neural networks, also known as twin neural networks, use shared weights to compare two input vectors and compute comparable outputs, with applications in face recognition, handwriting recognition, and text matchi…
AWS's open source Strands Robots SDK now integrates with Hugging Face Storage Buckets to enable a continuous data loop for robot learning, allowing users to record demonstrations, train policies, and deploy them to hardw…
Developer Manasvi Boineypally built Doc2Slides, a multi-agent AI pipeline that converts research papers into audience-tailored PowerPoint presentations. The system uses LangGraph to orchestrate five agents—parser, summar…
Natural Language Processing (NLP) is an AI field that enables computers to understand and generate human language, powering applications like chatbots, search engines, and content creation. Key techniques include tokeniz…
Mistral AI released OCR 4.1, a public preview of its optical character recognition service for Document AI, featuring native paragraph-level bounding box extraction, structural block labels, and block-level confidence sc…
Apple Machine Learning Research published a paper proposing that machine unlearning skip training points with negligible influence on model outputs, reporting computational savings of up to approximately 50 percent on re…
Independent researcher Taotuner has proposed a methodological framework called History-Dependent Dynamics (HDD) to disentangle history dependence, recurrence, self-reference, and self-modeling in dynamical systems. The f…