How well do explanation methods for machine-learning models work?
Researchers develop a way to test whether popular methods for understanding machine-learning models are working correctly.
Researchers develop a way to test whether popular methods for understanding machine-learning models are working correctly.
Spread out or bead up? A new process enables control over liquid-solid interfaces even with the most unlikely pairs of materials.
The more social behaviors a voice-user interface exhibits, the more likely people are to trust it, engage with it, and consider it to be competent.
New research suggests ways to optimize US climate policy design for a just energy transition.
Tiny Tides is an automated fast-flow instrument that can synthesize peptide-nucleic acids in a single shot.
Study reveals why some attempts to convert the greenhouse gas into fuel have failed, and offers possible solutions.
Prestigious grants will support full-time doctoral research abroad for the MIT architecture and urban planning students.
David Gamarnik has developed a new tool, the overlap gap property, for understanding computational problems that appear intractable.
Extra chromosome alters chromosomal conformation and DNA accessibility in neural progenitor cells; study establishes senescence as a potentially targetable mechanism for future treatment.
The targeted approach eliminated tumors in mice, with minimal side effects.
The discovery could offer a route to smaller, faster electronic devices.
With special treatment, minerals called zeolites — commonly found in cat litter — can efficiently remove the greenhouse gas from the air, researchers report.
MIT researchers are testing a simplified turbulence theory’s ability to model complex plasma phenomena using a novel machine-learning technique.
The new observations record a key crossover from classical to quantum behavior.
Embedding bacteria in soft material tests researchers’ theories.