Can large language models figure out the real world?
New test could help determine if AI systems that make accurate predictions in one area can understand it well enough to apply that ability to a different area.
Electrical engineering and computer science (EECS)
Researchers glimpse the inner workings of protein language models
A new approach can reveal the features AI models use to predict proteins that might make good drug or vaccine targets.
A new way to test how well AI systems classify text
As large language models increasingly dominate our everyday lives, new systems for checking their reliability are more important than ever.
Eco-driving measures could significantly reduce vehicle emissions
New research shows automatically controlling vehicle speeds to mitigate traffic at intersections can cut carbon emissions between 11 and 22 percent.
Helping data storage keep up with the AI revolution
Storage systems from Cloudian, co-founded by an MIT alumnus, are helping businesses feed data-hungry AI models and agents at scale.
MIT tool visualizes and edits “physically impossible” objects
By visualizing Escher-like optical illusions in 2.5 dimensions, the “Meschers” tool could help scientists understand physics-defying shapes and spark new designs.
New algorithms enable efficient machine learning with symmetric data
This new approach could lead to enhanced AI models for drug and materials discovery.
“FUTURE PHASES” showcases new frontiers in music technology and interactive performance
Groundbreaking MIT concert, featuring electronic and computer-generated music, was a part of the 2025 International Computer Music Conference.
Robot, know thyself: New vision-based system teaches machines to understand their bodies
Neural Jacobian Fields, developed by MIT CSAIL researchers, can learn to control any robot from a single camera, without any other sensors.
A new way to edit or generate images
MIT researchers found that special kinds of neural networks, called encoders or “tokenizers,” can do much more than previously realized.