Meta announced several research initiatives at its Open Source AI for Science Summit held last month at the company’s Menlo Park headquarters, according to a November 20 company post.
The company is expanding its Digital Brain Project, launched with the Rothschild Hospital Foundation and the University of Montreal. Meta increased the project’s target from 15,000 hours of de-identified brain data to 25,000 hours. The project funds researchers to collect and openly share brain data to help understand human intelligence and explore ways to improve neurological health, according to Meta.
Meta also released UMA Nitro, described as a faster version of its Universal Model for Atoms, which scientists use to run computer simulations for identifying promising materials and medicines before laboratory testing. In a study with Carnegie Mellon and UC Riverside, Meta’s UMA models identified multiple crystal forms of drug molecules without additional training data, according to the company. Understanding a drug’s crystal form helps researchers determine how a pill dissolves.
In work with the University of Chicago, Meta said UMA can simulate entire enzymes atom by atom, potentially helping scientists understand enzyme function and engineer versions for applications in medicine, sustainability and energy.
Meta also announced that its UMA Playground chemistry visualization tool is now available to high school chemistry teachers through Discovery Education, with classroom guides. The tool lets users build molecules, watch bonds break and form, and explore chemical reactions, according to the company.
Researchers from the University of Pittsburgh presented an assistive-robotics system built on Meta’s DINOv3 and SAM 3 models at the summit. The team plans to demonstrate a robot that can help a person eat and is experimenting with allowing users to give instructions in their own words, Meta said.