Large language models are biased. Can logic help save them?
MIT researchers trained logic-aware language models to reduce harmful stereotypes like gender and racial biases.
Topic
Electrical engineering and computer science (EECS)
Download RSS feed: News Articles / In the Media / Audio
Hari Balakrishnan awarded Marconi Prize
The prize is the top honor within the field of communications technology.
Report: CHIPS Act just the first step in addressing threats to US leadership in advanced computing
The Advanced Computing Users Survey, sampling sentiments from 120 top-tier universities, national labs, federal agencies, and private firms, finds the decline in America’s advanced computing lead spans many areas.
MIT-Takeda Program heads into fourth year with crop of 10 new projects
The program leverages MIT’s research expertise and Takeda’s industrial know-how for research in artificial intelligence and medicine.
Augmented reality headset enables users to see hidden objects
The device could help workers locate objects for fulfilling e-commerce orders or identify parts for assembling products.
Student-led conference charts the future of micro- and nanoscale research, reinforces scientific community
19th Microsystems Annual Research Conference reveals the next era of microsystems technologies, along with skiing and a dance party.
A new chip for decoding data transmissions demonstrates record-breaking energy efficiency
The chip, which can decipher any encoded signal, could enable lower-cost devices that perform better while requiring less hardware.
New chip for mobile devices knocks out unwanted signals
The receiver chip efficiently blocks signal interference that slows device performance and drains batteries.
A new way for quantum computing systems to keep their cool
A wireless technique enables a super-cold quantum computer to send and receive data without generating too much error-causing heat.
Nine from MIT named 2023 Sloan Research Fellows
Annual award honors early-career researchers for creativity, innovation, and research accomplishments.
MIT community members elected to the National Academy of Engineering for 2023
Seven researchers, along with 14 additional MIT alumni, are honored for significant contributions to engineering research, practice, and education.
Efficient technique improves machine-learning models’ reliability
The method enables a model to determine its confidence in a prediction, while using no additional data and far fewer computing resources than other methods.
Helping companies deploy AI models more responsibly
MIT spinout Verta offers tools to help companies introduce, monitor, and manage machine-learning models safely and at scale.
Scientists boost quantum signals while reducing noise
“Squeezing” noise over a broad frequency bandwidth in a quantum system could lead to faster and more accurate quantum measurements.
Solving a machine-learning mystery
A new study shows how large language models like GPT-3 can learn a new task from just a few examples, without the need for any new training data.