Accelerating AI tasks while preserving data security
The SecureLoop search tool efficiently identifies secure designs for hardware that can boost the performance of complex AI tasks, while requiring less energy.
The SecureLoop search tool efficiently identifies secure designs for hardware that can boost the performance of complex AI tasks, while requiring less energy.
Two studies find “self-supervised” models, which learn about their environment from unlabeled data, can show activity patterns similar to those of the mammalian brain.
Coauthors of a “Footwear Manifesto” report discuss survey findings that point to industry collaboration as a path to reducing waste in shoe manufacturing.
MIT computer scientists developed a way to calculate polygenic scores that makes them more accurate for people across diverse ancestries.
It’s not easy to parse young children’s words, but adults’ beliefs about what children want to communicate helps make it possible, a new study finds.
The low-cost FibeRobo, which is compatible with existing textile manufacturing techniques, could be used in adaptive performance wear or compression garments.
Using multiple observatories, astronomers directly detect tellurium in two merging neutron stars.
The vibrating platform could be useful for growing artificial muscles to power soft robots and testing therapies for neuromuscular diseases.
The fibers could help with testing treatments for nerve-related pain.
AI models that prioritize similarity falter when asked to design something completely new.
StructCode, developed by MIT CSAIL researchers, encodes machine-readable data in laser-cut objects by modifying their fabrication features.
Center for Ultracold Atoms gets funding boost to “punch through tough scientific barriers and see what's on the other side.”
A newly discovered type of electronic behavior could help with packing more data into magnetic memory devices.
When it comes to shaping political beliefs, MIT postdoc Chloe Wittenberg PhD ’23 finds video captivates, but might not beat text.