Microsoft Research unveils MatterSim AI materials, a new platform designed to accelerate materials science. It offers faster simulations and multi-task models for experimental synthesis and property prediction.
Microsoft Research unveils GridSFM electric grid model, a compact AI foundation model designed to rapidly predict AC optimal power flow. This innovation aims to boost efficiency and provide grid operators with better insights.
Microsoft Research introduces mimalloc memory allocator, an open-source, high-performance, and scalable solution for efficient memory management in modern applications and projects.
Microsoft Research has issued further notes on their recent work concerning AI delegation reliability. This update clarifies findings on how LLMs might affect documents in delegated workflows, emphasizing robust evaluation.
Microsoft Research introduces Vega zero-knowledge identity, a new approach for secure digital identity proofs. This system shares only necessary information, enhancing privacy in AI applications.
Microsoft Research has unveiled Data Formulator AI analytics 0.7, a new tool designed to streamline enterprise data workflows. It uses AI agents to help users explore, analyze, and visualize data for actionable insights.
Microsoft Research unveils MagenticLite, an innovative agentic system designed for small models. This development promises efficient AI performance, making advanced agentic AI small models accessible for everyday tasks across various platforms.
A new perspective from Microsoft Research suggests viewing AI as a trustworthy human intelligence extension, rather than a replacement. This approach aims to foster more reliable and ethical AI system development.