For Priya Donti, childhood trips to India were more than an opportunity to visit a large family. The biennial journey has stimulated her motivation to continue shaping her research and her teachings.
Contrasting with the Massachusetts family home, Donti – currently, the Silverman Family Career Development Professor in the Department of Electrical Engineering and Computer Science (EECS), MIT Schwarzman College of Computing and EECS, and the MIT Institute’s Chief Investigator (LIDS) were in the common position of the Information and Decision Systems (LIDS).
“It was quite clear to some extent in the issue of inequality ramping around the world,” Donti says. “From a young age, I knew I definitely wanted to deal with that issue.”
The motivation was even more exciting by a high school biology teacher. He focused on climate and sustainability.
“We’ve learned that climate change, this huge and important issue, exacerbates inequality,” Donti says. “It really stuck to me and set my belly on fire.”
So when Donty enrolled at Harvey Mad College, she thought that she would direct her energy to chemistry or materials science research to create the next generation of solar panels.
However, these plans were filted. Donti discovered work by British researchers who “fare in love” with computer science and who argued that artificial intelligence and machine learning were essential to integrating renewable energy into the power grid.
“It was the first time these two interests were gathered,” she says. “I got hooked and have been working on that topic ever since.”
Donti, who earned his PhD from Carnegie Mellon University, was able to design his degree to include computer science and public policy. In her research, she investigated the need for basic algorithms and tools that can manage power grids that rely heavily on renewable energy.
“We wanted to reach out to develop these algorithms and toolkits by creating new machine learning technologies based on computer science,” she says. “But I wanted to make sure my work was done in a real energy system domain and people in that domain to provide what I actually need.
While Donti was completing her PhD, she co-founded a nonprofit called Climate Change AI. Her aim was to support a community of people involved in climate and sustainability, “they are computer scientists, academics, practitioners, or policy makers,” and to access resources, connections, education, “to help them along that journey.”
“In the climate field, there must be experts in specific climate-related sectors, experts in various technical and social science toolkits, problem owners, users affected, policy makers who know the regulations, that is, scalable impacts on the ground.”
When Donti came to MIT in September 2023, it was not surprising that she was drawn out by an initiative directing the application of computer science to the biggest issues in society, particularly the current threat to the health of the planet.
“We really think about where technology is a place where Horizon has much longer courage and how technology, society and policy all have to work together,” Donti says. “Technology is not just one in a year’s context, it’s not monetisable.”
Her work incorporates the physics and hard constraints of power systems that employ renewable energy for better prediction, optimization and control using deep learning models.
“Machine learning is already very widely used in solar prediction and more, and is a prerequisite for balancing the management and power grid,” she says. “My focus is how do you improve the algorithms to actually balance the power grid in the face of renewable energy in the range of time-varying?”
Among Donti’s breakthroughs is a promising solution where power grid operators can optimize costs, taking into account the actual physical reality of the grid, rather than relying on approximations. The solution has not yet been deployed, but it appears to work 10 times faster than previous technology and far cheaper, attracting the attention of grid operators.
Another technology she is developing works to provide data that can be used to train machine learning systems for power system optimization. Generally, many of the data associated with a system are unique or due to security concerns. Donti and her research group are working on creating synthetic data and benchmarks, Donti said, in order to make the power system more efficient.
“The question is, can they take the dataset to the point so that they are difficult enough to drive progress?”
For her efforts, Donti was awarded a Computational Science Alumni Fellowship from the U.S. Department of Energy and a NSF Graduate Studies Fellowship. She was recognized as part of MIT Technology ReviewList of “35 Innovators Under 35” for 2021 and Vox’s 2023 Future Perfect 50.
Next spring, Donti will co-teach a class called AI for Climate Action with EECS Assistant Professor Sarah Bealy.
“We’re all very excited about it,” Donti says.
Coming to MIT, Donti said, “I knew there was an ecosystem of people who really cared about not just the indicators of success like publications and citations, but also the impact of work on society.”
