Machine learning news — deep learning, reinforcement learning, neural architecture search, diffusion models, and new ML frameworks and libraries.
An international research team found that large language models integrated false statements into their outputs even when those claims were explicitly labeled as false in training data. The researchers seeded fine-tuning …
A solo developer launched Agenlus, a browser-based reinforcement learning training platform designed to make RL accessible without requiring expensive hardware or installations. The platform allows users to train agents,…
Azercell Telecom LLC, Azerbaijan's leading telecommunications provider, trained an Azerbaijani large language model on Amazon SageMaker AI in collaboration with the AWS Generative AI Innovation Center. The six-week proje…
Researchers from Meta, Google, the University of Maryland, University of Virginia, Washington University in St. Louis, and the University of North Carolina developed AutoTTS, a framework that automatically discovers opti…
A developer benchmarked seven Retrieval-Augmented Generation (RAG) pipelines over 20 hours of compute time, then open-sourced the resulting system, SEQUOIA. Another engineer argued that teams should prioritize engineerin…
A developer benchmarked seven RAG configurations against real-world banking documents, technical manuals, and internal wikis, finding that Twitter-hyped "graph RAG" approaches like LightRAG performed poorly. The engineer…
A developer built an embedding-based classifier using logistic regression to triage personal emails into priority, FYI, and ignore categories. The client-side application, which runs in a browser and uses OpenAI for embe…
Amazon Web Services released a solution for building custom portals that embed Amazon SageMaker AI MLflow Apps, allowing machine learning teams to access experiment tracking through a single bookmarkable URL without pres…
AWS announced a solution for organizations to integrate Amazon SageMaker MLflow with existing enterprise systems through a Flask-based REST API proxy. The proxy service provides HTTPS access to SageMaker MLflow without r…
LangChain and AWS have released a framework for evaluating AI agents using LangSmith on Amazon Bedrock, addressing the challenge of validating non-deterministic, multi-step agent behavior before production. The approach …
GPU utilization rates on major clouds average just 5 percent, meaning most paid GPU hours produce no useful work, while egress fees of up to $0.12 per GB and data gravity lock-in add hidden costs that can dwarf the headl…
A growing portion of the internet is now generated by AI, creating a recursive training loop where models are increasingly trained on their own outputs rather than on original human data. This shift is causing a subtle b…
Researchers have developed SIA, a self-improving AI system that updates both its own software scaffolding and internal model weights without human intervention, combining two previously separate approaches to AI improvem…
Researchers have identified a structural similarity between three distinct AI safety phenomena: negation neglect, inoculation prompting non-robustness, and backdoor non-robustness. In each case, training a model with a s…
Canonical and Google Cloud announced the availability of certified Ubuntu images for Cloud TPU virtual machines on Google Cloud, with the operating system now preinstalled by default when setting up the VMs. The integrat…
A developer built a 3.7 kB edge AI smoke detector in pure C, using a neural network that runs on ultra-low-power microcontrollers like the ESP32. The model monitors 12 environmental variables in parallel to detect non-li…
Mixture of Experts (MoE) models are reshaping the economics of open-model deployments by reducing GPU inference costs and altering serving stack requirements. The shift toward MoE architectures in 2026 forces engineering…
Amazon Bedrock AgentCore now supports dataset management for agent evaluation, enabling developers to create versioned test fixtures with inputs, expected outputs, assertions, and tool sequences. The feature allows teams…
Researchers have developed Paris 2.0, the first video generation model pre-trained through decentralized computation across heterogeneous GPUs, eliminating the need for a monolithic cluster. The model achieved a 2x impro…
Hexo AI has released SIA, an open-source self-improving AI system that can recursively update its own code, model weights, and memory layer based on performance feedback. The system outperforms specialized agents on benc…