A new way to increase the capabilities of large language models
MIT-IBM Watson AI Lab researchers developed an expressive architecture that provides better state tracking and sequential reasoning in LLMs over long texts.
A “scientific sandbox” lets researchers explore the evolution of vision systems
The AI-powered tool could inform the design of better sensors and cameras for robots or autonomous vehicles.
New study suggests a way to rejuvenate the immune system
Stimulating the liver to produce some of the signals of the thymus can reverse age-related declines in T-cell populations and enhance response to vaccination.
How cement “breathes in” and stores millions of tons of CO₂ a year
New analysis provides the first national, bottom-up estimate of cement’s natural carbon dioxide uptake across buildings and infrastructure.
Introducing the Minerals Stewardship Consortium at MIT
The consortium convenes industry, academia, and policy leaders to navigate competing demands and reimagine materials supply.
A new immunotherapy approach could work for many types of cancer
Using new molecules that block an immune checkpoint, researchers showed they could stimulate a strong anti-tumor immune response.
“Robot, make me a chair”
An AI-driven system lets users design and build simple, multicomponent objects by describing them with words.
RNA editing study finds many ways for neurons to diversify
Tracking how fruit fly motor neurons edit their RNA, neurobiologists cataloged hundreds of target sites and varying editing rates, finding many edits altered communication- and function-related proteins.
What makes a good proton conductor?
MIT researchers found a way to predict how efficiently materials can transport protons in clean energy devices and other advanced technologies.
Deep-learning model predicts how fruit flies form, cell by cell
The approach could apply to more complex tissues and organs, helping researchers to identify early signs of disease.
MIT HEALS leadership charts a bold path for convergence in health and life sciences
Angela Koehler, Iain Cheeseman, and Katharina Ribbeck are shaping the collaborative as a platform for transformative research, translation, and talent development across MIT.
Enabling small language models to solve complex reasoning tasks
The “self-steering” DisCIPL system directs small models to work together on tasks with constraints, like itinerary planning and budgeting.
Making clean energy investments more successful
Tools for forecasting and modeling technological improvements and the impacts of policy decisions can result in more effective and impactful decision-making.
New method improves the reliability of statistical estimations
The technique can help scientists in economics, public health, and other fields understand whether to trust the results of their experiments.
MIT researchers find new immunotherapeutic targets for glioblastoma
A study profiling antigens presented on immune and tumor cells in co-culture points to new strategies for attacking a treatment-resistant and deadly brain cancer.