|
Activity Number:
|
504
- Computational Challenges in Modern Statistical Inference
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 11, 2022 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
IMS
|
|
Abstract #323890
|
|
|
Title:
|
Low-Rank Matrix Estimation with Groupwise Heteroskedasticity
|
|
Author(s):
|
Galen Reeves*
|
|
Companies:
|
Duke University
|
|
Keywords:
|
|
|
Abstract:
|
I will discuss recent results describing the fundamental limits of recovery for a broad class of problems involving heterogeneous pairwise observations. The key idea is to show that many of these problems can be seen as a special case of the matrix tensor product model. An application of these results shows that recently proposed methods based on applying principal component analysis to weighted combinations of the data are optimal in some settings but suboptimal in others.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2022 program
|
