Like human brains, large language models reason about diverse data in a general way
A new study shows LLMs represent different data types based on their underlying meaning and reason about data in their dominant language.
Search Stories
Engineers enable a drone to determine its position in the dark and indoors
A new low-power system using radio frequency waves takes a major step toward autonomous, indoor drone navigation.
To keep hardware safe, cut out the code’s clues
New “Oreo” method from MIT CSAIL researchers removes footprints that reveal where code is stored before a hacker can see them.
Creating smart buildings with privacy-first sensors
Butlr, founded by former Media Lab researchers, uses insights from thermal sensors to make buildings safe and efficient.
Mapping mRNA through its life cycle within a cell
Xiao Wang’s studies of how and where RNA is translated could lead to the development of better RNA therapeutics and vaccines.
Evelina Fedorenko receives Troland Award from National Academy of Sciences
Cognitive neuroscientist is recognized for her groundbreaking discoveries about the brain’s language system.
MIT students' works redefine human-AI collaboration
Projects from MIT course 4.043/4.044 (Interaction Intelligence) were presented at NeurIPS, showing how AI transforms creativity, education, and interaction in unexpected ways.
How good old mud can lower building costs
Builders pour concrete into temporary molds called formwork. MIT researchers invented a way to make these structures out of on-site soil.
A platform to expedite clean energy projects
Station A, founded by MIT alumni, makes the process of buying clean energy simple for property owners.
Toward video generative models of the molecular world
Starting with a single frame in a simulation, a new system uses generative AI to emulate the dynamics of molecules, connecting static molecular structures and developing blurry pictures into videos.
New general law governs fracture energy of networks across materials and length scales
Findings reported by MIT researchers may have significant implications in material design.
The multifaceted challenge of powering AI
Providing electricity to power-hungry data centers is stressing grids, raising prices for consumers, and slowing the transition to clean energy.
New computational chemistry techniques accelerate the prediction of molecules and materials
With their recently-developed neural network architecture, MIT researchers can wring more information out of electronic structure calculations.
Minimizing the carbon footprint of bridges and other structures
MAD Design Fellow Zane Schemmer writes algorithms that optimize overall function, minimize carbon footprint, and produce a manufacturable design.
The regions racing to become the “Silicon Valley” of an aging world
New book from MIT AgeLab researchers applies the concept of innovation clusters to global population aging.