Microsoft Research Focus – Edition #97
This digest provides a useful overview of Microsoft's ongoing research into improving AI agent reliability, enterprise AI adoption patterns, and scientific foundation models.
What happened
Edition #97 of Microsoft Research Focus highlights several recent AI studies and releases from the company, including the ForkPilot framework for reducing agent processing demands and the CABRA benchmark for identifying why coding agents fail. The roundup also covers TeleTune for training computer-use agents from activity logs, an analysis of enterprise Copilot usage, the Shiksha Copilot study for teachers in India, a machine-learning framework for extended-range weather forecasting, and the public preview of the BioEmu2 biomolecular foundation model.
Source
Microsoft Research
Microsoft Research Focus – Edition #97
Get a brief picked for you
Sign in to follow topics and companies, save stories and search the archive.
Sign in