News & features
Research Focus: Week of March 24, 2025
In this issue, we examine a new conversation segmentation method that delivers more coherent and personalized agent conversation, and we review efforts to improve MLLMs’ understanding of geologic maps. Check out the latest research and other updates.
Claimify: Extracting high-quality claims from language model outputs
| Dasha Metropolitansky
Claimify, created by Microsoft Research, is a novel LLM-based claim-extraction method that outperforms prior solutions to produce more accurate, comprehensive, and substantiated claims from LLM outputs.
In the news | Becker’s Healthcare
Jim Weinstein named among great leaders in healthcare by Becker’s Hospital Review
Dr. Weinstein is a leading advocate for population health, patient agency, and healthcare transformation.
Semantic Telemetry: Understanding how users interact with AI systems
| Amber Hoak, Scott Counts, Kate Lytvynets, David Tittsworth, Siddharth Suri, Ben Cutler, and Weiwei Yang
AI interactions can be iterative and complex. Learn how the Semantic Telemetry project at Microsoft Research is developing a new data science approach to understand human-AI interactions and their value.
In the news | BBC
Tech Life: The doctor will see you now
Find out how holoportation 3D telemedicine technology is helping patients in Ghana in this BBC TechLife podcast.
Jianwei Yang introduces Magma, a new multimodal agentic foundation model designed for UI navigation in digital environments and robotics manipulation in physical settings.
In the news | TheSequence
One of the Best Agent Frameworks in the Market Just Got Way Better
AutoGen has undergone significant evolution since its inception, driven by the need for more efficient, flexible, and scalable agentic AI systems. The release of AutoGen v0.4 introduces a fundamental architectural shift, addressing prior inefficiencies and enhancing its capabilities.
In the news | Social Media Fails Many Users. Experts Have an Idea to Fix It
Remaking “Prosocial” Media
E. Glen Weyl and his coauthors have proposed a new “prosocial” structure for social media, where posts are flagged with the context of where they are socially accepted and users can navigate to see opposing perspectives on divisive issues.
Research Focus: Week of January 27, 2025
In this issue: A new approach to multimodal pretraining for remote sensing; Managed-retention memory for the AI era; Improving detection of macular telangiectasia type 2; Generalizing symbolic automata.