For Priya Donty, childhood trips were more than the opportunity to meet an extended family. Biennial visits raised an inspiration in them that shaped their research and teaching.
Unlike his family home in Massachusetts, Donty – which is now the Silverman Family Career Development Professor in the Department of Electrical Engineering and Computer Science (EECS), is a shared position between MIT Schwarzeman College of Computing and EECS, and is a prominent internalist in the MIT laboratory for the information and decision system (LIDS) – a major non -attendant in the MIT laboratory for the information and decision system (LIDS). He was influenced by inequalities.
Donty says, “It was very clear to me to what extent inequality is a big issue worldwide.” “From a young age, I knew that I definitely wanted to solve that issue.”
The inspiration was further promoted by a high school biology teacher, which focused his class on climate and stability.
“We learned that climate change, this is a very big, important issue, will increase inequality,” says Donty. “He really dominated me and set my stomach on fire.”
Therefore, when Donty enrolled in Harvey Mad College, she thought she would put her energy towards the study of chemistry or material science to create the next generation solar panels.
However, those plans collapsed. Donty “fell in love with computer science”, and then discovered the work of researchers in the United Kingdom who were arguing that artificial intelligence and machine learning would be necessary to help integrate renewable energy into the power grid.
She says, “This was the first time I saw those two interests coming together.” “I joined it and have been working on that subject since then.”
While studying PhD at the University of Carnegie Melan, Donty was capable of designing his degree to include computer science and public policy. In his research, he discovered the need for fundamental algorithms and equipment that could manage the power grid dependent on renewable energy on a large scale.
She says, “I wanted to shake hands in developing those algorithms and tool kits by creating new machine learning techniques based on computer science.” “But I wanted to make sure that the way I was working was based on both working with the real energy system area and people in that field” so that what really needed could be provided.
While Donty was working on her PhD, she co-established a non-profit organization called Climate Change AI. She says, her aim was to help the community of people involved in climate and stability – “whether they are computer scientists, academics, businessmen, or policy makers” – coming together and accessing resources, connections and education “to help them in that journey.”
“In the climate zone,” she says, “you especially have experts in the fields related to climate change, specialists of various technical and social science tool kits, problem owners, affected users, policy makers who know the rules – all these – to impact the ground level.”
When the Donty came to MIT in September 2023, it was not surprising that she was attracted to its initiative, directing the application of computer science to the biggest problems of society, especially the right threat to the health of the planet.
Donty says, “We are really thinking about where the technology has a long-wide effect and how the technology, society and policy will all work together.” “Technology is not only one-and-a-year-old and mudilable in terms of one year.”
Their work uses intensive teaching models to include physics and difficult obstacles of electrical energy systems that use renewable energy for better forecasts, adaptation and control.
She says, “Machine learning is already widely used for things such as solar energy forecasts, which is a condition for the management and balance of the power grid.” “My focus is on how do you really improve the algorithm to balance the power grid in the event of changing renewable energy from time to time?”
Power grid in Donty’s successes is a promising solution to be able to optimize the cost keeping in mind the actual physical realities of the grid rather than relying on estimates. Although the solution has not been deployed yet, it appears that it works 10 times faster and more cheaply than in previous technologies, and it has attracted the attention of grid operators.
She is developing another technology that works to provide data that can be used in training of machine learning systems for power system optimization. In general, most of the data related to the system is private, either due to ownership or safety concerns. Donty and his research groups are working to create synthetic data and benchmarks, which Donty says, can help “some underlying problems” in making power systems more efficient “.
“The question is,” Donty says, “Can we bring our dataset to such a point that they are hard enough to pursue progress?”
For his efforts, Donty has been awarded the US Department of Energy Computer Science Graduate Fellowship and NSF Graduate Research Fellowship. He was identified as part of MIT technology reviewThe 2021 list of “35 Innovators Under 35” and Vax’s 2023 “Future Perfect 50.”
In the next spring, Donty EECS will co-educate an orbit called AI for climate action with Assistant Professor Sara Berry, which is AI for biodiversity and ecosystem, and Abigail Bodner, EECS and Earth are assistant professors in the departments whose focus is AI for Jalwas and Earth Sciences.
Donty says, “We are all very excited about it.”
Coming to the MIT, Donty says, “I knew that there would be an ecosystem of people who actually cares about success metrics such as publications and quotes calculations, but also about the impact of our work on society.”
