[Theory Lunch 03/26] Yu Cheng: High-Dimensional Robust Statistics: Faster Algorithms and Optimization Landscape
16 views
Skip to first unread message
Haoming Li
Mar 22, 2021, 4:07:56 AM3/22/21
to usc-theo...@googlegroups.com, usc-t...@googlegroups.com
USC CS Theory Lunch:
High-Dimensional Robust Statistics: Faster Algorithms and Optimization Landscape
High-Dimensional Robust Statistics: Faster Algorithms and Optimization Landscape
Speaker: Yu Cheng (https://homepages.math.uic.edu/~yucheng/)
Time: 03/26/21 11:45am PST
Location: https://usc.zoom.us/j/94386654763
Abstract:
We study the fundamental problem of high-dimensional robust estimation where a constant fraction of the samples are adversarially corrupted. Recent work gave the first polynomial-time robust algorithms for basic statistical problems with dimension-independent error guarantees.
In this talk, we will discuss several recent results in high-dimensional robust statistics, focusing on (1) designing robust estimators that run as fast as their non-robust counterparts, and (2) exploring the optimization landscape of more direct non-convex formulations of robust estimation.
Most of the talk is based on joint work with Ilias Diakonikolas, Rong Ge, and Mahdi Soltanolkotabi.
Time: 03/26/21 11:45am PST
Location: https://usc.zoom.us/j/94386654763
Abstract:
We study the fundamental problem of high-dimensional robust estimation where a constant fraction of the samples are adversarially corrupted. Recent work gave the first polynomial-time robust algorithms for basic statistical problems with dimension-independent error guarantees.
In this talk, we will discuss several recent results in high-dimensional robust statistics, focusing on (1) designing robust estimators that run as fast as their non-robust counterparts, and (2) exploring the optimization landscape of more direct non-convex formulations of robust estimation.
Most of the talk is based on joint work with Ilias Diakonikolas, Rong Ge, and Mahdi Soltanolkotabi.
Reply all
Reply to author
Forward
0 new messages