Doing machine learning the right way

MIT Professor Aleksander Madry strives to build machine-learning models that are more reliable, understandable and robust.

Cryptographic ID tag developed to stop supply chain counterfeiting

MIT researchers have invented a cryptographic ID tag that’s small enough to fit on virtually any product and verify its authenticity, which could help stop supply chain counterfeiting.

Soft robotic arm uses flexible sensors to understand its position

MIT researchers have developed flexible sensors and an artificial intelligence model that tell deformable robots how their bodies are positioned in a 3D environment.

Machine-learning tool predicts code’s speed on computer chips

MIT researchers have invented a machine-learning tool called Ithemal, which predicts how fast computer chips will execute code from various applications.  

Design and control robots with stretchy, flexible bodies

MIT researchers have found a way top optimize soft robots to perform specific tasks with a new model that make their bodies flexible.

Automated system generates robotic actuators

MIT researchers have developed an automated system that designs and 3-D prints robotic actuators, which are optimized and automatically created, which is almost impossible for a human to do.

Particle simulator developed to help robots better predict and grasp objects

MIT researchers have developed a learning-based particle simulator, could give industrial robots a more refined touch and help them make predictions about interacting with solid objects and liquids.

Researchers develop architecture to reduce data center energy and space constraints

MIT researchers have developer a flash-storage system designed to cut the energy and physical space required to store and manage data in data centers by half.