Training LLMs to self-detoxify their language
A new method from the MIT-IBM Watson AI Lab helps large language models to steer their own responses toward safer, more ethical, value-aligned outputs.
New method efficiently safeguards sensitive AI training data
The approach maintains an AI model’s accuracy while ensuring attackers can’t extract secret information.
Researchers teach LLMs to solve complex planning challenges
This new framework leverages a model’s reasoning abilities to create a “smart assistant” that finds the optimal solution to multistep problems.
Robotic helper making mistakes? Just nudge it in the right direction
New research could allow a person to correct a robot’s actions in real-time, using the kind of feedback they’d give another human.
Letterlocking: A new look at a centuries-old practice
A first history of the document security technology, co-authored by MIT Libraries’ Jana Dambrogio, provides new tools for interdisciplinary research.
Sometimes, when competitors collaborate, everybody wins
Engineers developed a planning tool that can help independent entities decide when they should invest in joint projects.
Rooftop panels, EV chargers, and smart thermostats could chip in to boost power grid resilience
MIT engineers propose a new “local electricity market” to tap into the power potential of homeowners’ grid-edge devices.
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.
MIT spinout maps the body’s metabolites to uncover the hidden drivers of disease
ReviveMed uses AI to gather large-scale data on metabolites — molecules like lipids, cholesterol, and sugar — to match patients with therapeutics.
AI model deciphers the code in proteins that tells them where to go
Whitehead Institute and CSAIL researchers created a machine-learning model to predict and generate protein localization, with implications for understanding and remedying disease.
Puzzling out climate change
Accenture Fellow Shreyaa Raghavan applies machine learning and optimization methods to explore ways to reduce transportation sector emissions.
Streamlining data collection for improved salmon population management
Assistant Professor Sara Beery is using automation to improve monitoring of migrating salmon in the Pacific Northwest.
3 Questions: Modeling adversarial intelligence to exploit AI’s security vulnerabilities
MIT CSAIL Principal Research Scientist Una-May O’Reilly discusses how she develops agents that reveal AI models’ security weaknesses before hackers do.
New training approach could help AI agents perform better in uncertain conditions
Sometimes, it might be better to train a robot in an environment that’s different from the one where it will be deployed.
Expanding robot perception
Associate Professor Luca Carlone is working to give robots a more human-like awareness of their environment.