[AIRG] Robust regression via hard thresholding

[XIAOMIN ZHANG <xzhang682@xxxxxxxx>]

Hi AIRG,

I will talk about robust regression this Wednesday, which does much better than standard regression at handling noisy data, corrupted data, or outliers. Robust regression via convex methods may not be consistent. However, non-convex methods can be consistent. I will focus on hard thresholding for robust regression. 

References are the following papers:

1. https://arxiv.org/pdf/1506.02428.pdf

2. https://papers.nips.cc/paper/6806-consistent-robust-regression.pdf

See you at 4pm Wednesday, CS 4310.,

Xiaomin