Teaching AI models what they don’t know
A team of MIT researchers founded Themis AI to quantify AI model uncertainty and address knowledge gaps.
Algorithms
Teaching AI models the broad strokes to sketch more like humans do
SketchAgent, a drawing system developed by MIT CSAIL researchers, sketches up concepts stroke-by-stroke, teaching language models to visually express concepts on their own and collaborate with humans.
An anomaly detection framework anyone can use
PhD student Sarah Alnegheimish wants to make machine learning systems accessible.
AI learns how vision and sound are connected, without human intervention
This new machine-learning model can match corresponding audio and visual data, which could someday help robots interact in the real world.
Learning how to predict rare kinds of failures
Researchers are developing algorithms to predict failures when automation meets the real world in areas like air traffic scheduling or autonomous vehicles.
The sweet taste of a new idea
Sendhil Mullainathan brings a lifetime of unique perspectives to research in behavioral economics and machine learning.
New tool evaluates progress in reinforcement learning
“IntersectionZoo,” a benchmarking tool, uses a real-world traffic problem to test progress in deep reinforcement learning algorithms.
Novel AI model inspired by neural dynamics from the brain
New type of “state-space model” leverages principles of harmonic oscillators.
Designing a new way to optimize complex coordinated systems
Using diagrams to represent interactions in multipart systems can provide a faster way to design software improvements.
“Periodic table of machine learning” could fuel AI discovery
Researchers have created a unifying framework that can help scientists combine existing ideas to improve AI models or create new ones.