Method prevents an AI model from being overconfident about wrong answers
More efficient than other approaches, the “Thermometer” technique could help someone know when they should trust a large language model.
More efficient than other approaches, the “Thermometer” technique could help someone know when they should trust a large language model.
A mathematical method, validated with experimental data, provides a fast, reliable, and minimally invasive way of determining how to treat critical blood pressure changes during surgery or intensive care.
Through MISTI’s Imperial College London Exchange, students experience AeroAstro, MIT, and the beauty of New England.
The company that brought you no-stick toothpaste is moving into the medical space, with a lubricant for ostomy pouches and other products that could improve millions of lives.
United Kingdom Supply Chain and Logistics Excellence Centre (UK SCALE) joins prestigious international network to advance global supply chain and logistics innovation.
Ultrathin material whose properties “already meet or exceed industry standards” enables superfast switching, extreme durability.
Drone company founders with MIT Advanced Study Program roots seek to bring aerial delivery to the mainstream.
MIT engineers have developed a fast and sustainable method for producing hydrogen fuel using aluminum, saltwater, and coffee grounds.
Domitilla Del Vecchio and Themis Sapsis of MechE and Mehrdad Jazayeri of BCS will each receive up to $3 million for blue-sky research.
MIT historian Tristan Brown describes how China’s feng shui legacy can help with confronting today’s climate challenges.
The effort to accelerate climate work at the Institute adds to its leadership team.
Genomics and lab studies reveal numerous findings, including a key role for Reelin amid neuronal vulnerability, and for choline and antioxidants in sustaining cognition.
Introducing structured randomization into decisions based on machine-learning model predictions can address inherent uncertainties while maintaining efficiency.
MAIA is a multimodal agent that can iteratively design experiments to better understand various components of AI systems.
Analysis and materials identified by MIT engineers could lead to more energy-efficient fuel cells, electrolyzers, batteries, or computing devices.