UCLA Computer Scientist Receives NSF Funding to Train AI for Cosmic Data Analysis
Baharan Mirzasoleiman, an assistant professor of computer science at the UCLA Samueli School of Engineering, has received $550,000 from the National Science Foundation and the Simons Foundation to train machine-learning algorithms to process astronomical data.
Accurate data analysis powered by machine learning could greatly accelerate time-consuming processes such as modeling the complex chemical reactions inside stars. With less time sunk into processing data to generate sophisticated models, researchers can devote more effort to uncovering new insights about the cosmos.
“Currently, the volume of the astronomy data is so large that it makes data processing prohibitively expensive,” said Mirzasoleiman, who leads the BigML research group at UCLA Samueli. “My role will be to design algorithms that can efficiently train foundation machine learning models by extracting the information from massive amounts of astronomy data.”
The funding for Mirzasoleiman’s project is part of a larger grant that established the NSF-Simons AI Institute for Cosmic Origins (CosmicAI). Researchers from the University of Texas at Austin will lead the five-year initiative comprising a cross-disciplinary team of researchers from several universities.
“Astronomy has incredibly rich and open datasets and is poised for more deep and profound inquiry,” said Simons Foundation president David Spergel in the initiative announcement. “AI offers novel tools that can use this data both to produce transformative results and to develop tools that can have impact in other fields.”
With AI-assisted data processing spearheaded by Mirzasoleiman, astronomers and other researchers can more efficiently gain a deep understanding of the wealth of data available to them.
The eventual goal of NSF-Simons CosmicAI is to democratize access to astronomical data, both through providing AI-powered, data-sourced answers to scientific queries and through training early-career researchers and students.
Mirzasoleiman, who joined UCLA Samueli in 2020, focuses her research on improving the quality of big data by developing theoretically rigorous methods to select the most beneficial data for efficient and robust learning. Her research aims to address sustainability, reliability and efficiency of machine learning covering data from a wide range of applications, including medical diagnosis and environment sensing.