JUMP TO SECTION
Accelerated Materials
Accelerated Systems
Education
Research Accomplishments
We need to develop new materials 10–100x faster, to meet pressing climate and social needs. Our team explores how automation, machine learning, and computing can combine to realize this need. We’ve customized over twenty machine-learning algorithms to learn new physics, optimize processes faster, characterize samples faster, and drive autonomous labs in closed-loop cycles of learning.
At MIT and at SMART, we focus on photovoltaics (solar energy). Through our collaborations, we apply our methods to other materials of societal benefit, including sustainable polymers, tunable nanotechnology, and more. We’ve designed twenty new materials using machine learning (probably more, by the time you’re reading this). Read our latest papers and LinkedIn summaries for more.
We accelerate novel systems development with machine learning. Collaborative examples include generative design and optimization of solar cell devices for custom operating environments, solar-powered desalination units, optimization of solar manufacturing equipment, and more. We apply our methods broadly, to accelerate the development of new systems that include storage, water, energy, communications, and sensing elements.
For individuals to up-skill, access to open-source learning materials is needed. The Accelerated Materials Development YouTube channel shares success stories of machine learning applied to materials science problems, complete with Jupyter notebooks, datasets, and code. The videos were recorded at a workshop series for engineers and scientists.
Prior to this, Prof. Buonassisi authored an OpenCourseware course, Fundamentals of Photovoltaics, which received over 70k views, and for which he received the prestigious Everett Moore Baker Award for Excellence in Undergraduate Teaching at MIT.
Research 1 - 10 SHOW MORE