Researchers develop a way to test whether popular methods for understanding machine-learning models are working correctly.

Deep-learning methods confidently recognize images that are nonsense, a potential problem for medical and autonomous-driving decisions.

The system could help physicians select the least risky treatments in urgent situations, such as treating sepsis.

A new “common-sense” approach to computer vision enables artificial intelligence that interprets scenes more accurately than other systems do.

“Evolution Gym” is a large-scale benchmark for co-optimizing the design and control of soft robots that takes inspiration from nature and evolutionary processes.

The new machine-learning system can generate a 3D scene from an image about 15,000 times faster than other methods.

A new AI-powered, virtual platform uses real-world physics to simulate a rich and interactive audio-visual environment, enabling human and robotic learning, training, and experimental studies.

A deep learning model rapidly predicts the 3D shapes of drug-like molecules, which could accelerate the process of discovering new medicines.

A new machine-learning model could enable robots to understand interactions in the world in the way humans do.

New work on linear-probing hash tables from MIT CSAIL could lead to more efficient data storage and retrieval in computers.