News

MIT CSAIL researchers combined generative AI and a physics simulation engine to create a machine that outjumped a robot designed by a human (Credit: Researchers photographed by Dan McDonald and image collaged by Alex Shipps/MIT CSAIL using assets from the researchers).

Using generative AI to help robots jump higher and land better

An AI-generated robot learns to understand its own body (Credit: Alex Shipps/MIT CSAIL using Midjourney).

Robots that know themselves: MIT’s vision-based system teaches machines to understand their bodies

A small molecule binds to an OX2 protein. The new foundation model Boltz-2, developed by researchers at MIT and Recursion, achieves state-of-the-art performance in protein binding affinity prediction (Image: Courtesy of the researchers).

MIT releases breakthrough protein-binding affinity model, expanding role of AI in drug discovery

Spotlighted News

Using generative AI to help robots jump higher and land better

Robots that know themselves: MIT’s vision-based system teaches machines to understand their bodies

MIT releases breakthrough protein-binding affinity model, expanding role of AI in drug discovery