Michał Dereziński receives NSF CAREER Award to support next generation of data science algorithms

Dereziński aims to develop improved algorithms for data science applications.
Prof. Michał Dereziński
Prof. Michał Dereziński

Michał Dereziński, assistant professor of computer science and engineering at the University of Michigan, has received a National Science Foundation (NSF) CAREER Award in support of his research to design and deploy improved linear algebra algorithms for computational data science. His project, titled “Leveraging Randomization and Structure in Computational Linear Algebra for Data Science,” will help promote the development of more efficient algorithms as well as their adoption in data science applications.

The NSF CAREER Award is a distinguished honor given to early-career faculty across all science and engineering disciplines based on their research and education excellence. The award recognizes and supports young scientists and engineers who “have the potential to serve as academic role models in research and education and to lead advances in the mission of their department of organization.”

Data science has come to play an increasingly important role in modern society, providing key information to guide our understanding and decision making surrounding healthcare, the environment, and beyond. The ability to glean meaningful information from data in these and countless other areas relies on algorithms, which use linear algebraic objects such as matrices to represent data.

While linear algebra algorithms form the foundation of data science, there is a significant disparity between the theoretical development and study of these algorithms and their adoption in real-world applications.

Through his research, Dereziński aims to bridge this theory-practice gap by bringing together efforts to both design improved randomization algorithms and deploy them across computational data science. To accomplish this, he will pursue novel randomized approaches for designing efficient algorithms for ubiquitous matrix problems, including matrix multiplication and low-rank approximation. This research will contribute to the development of linear algebra algorithms that are able to preserve data structure, which is critical to supporting their use in data science applications.

Another core focus of Dereziński’s project is integrating algorithmic foundations and data science skills into his teaching. He also plans to join forces with the U-M Engineering Pathways program, which targets underresourced high schools across Michigan, to support educational outreach in these areas.

An expert in computer theory, Dereziński joined the CSE faculty in 2021. He earned his PhD in computer science at the University of California, Santa Cruz, in 2018, after which he was a postdoctoral fellow in the Department of Statistics at the University of California, Berkeley, and a research fellow at the Simons Institute for the Theory of Computing. His research focuses on the theoretical foundations of randomized algorithms for machine learning and data science. He has published several papers in this area, one of which won Best Paper Award at NeurIPS 2020.