YINS Seminar Archives: Paxton Turner (Dec. 4, 2020)

YINS Seminar Archives: Paxton Turner (Dec. 4, 2020)

Talk Summary: 

“A Statistical Perspective on Coresets”

Speaker: Paxton Turner, Department of Mathematics at MIT

Coresets have emerged as a powerful tool to summarize data by selecting a small subset of the original observations while retaining most of its information. This approach has led to significant computational speedups, but the statistical performance of procedures run on coresets is largely unexplored. I will introduce a statistical framework to study coresets and focus on the task of nonparametric density estimation. Then I will characterize the statistical performance of coreset-based estimators and show that the practical coreset kernel density estimators are near-minimax optimal. 

This presentation was a part of the YINS Speaker Series and was presented on Friday, December 4, 2020.

Speaker: 
Paxton Turner (MIT)
Bio: 

Paxton Turner is a fifth year Ph.D. student in the department of mathematics at MIT, advised by Philippe Rigollet. His research interests span statistics, algorithms, and theoretical aspects of machine learning. Before coming to MIT, he did his undergraduate studies in mathematics at Louisiana State University.