Subspace Clustering with Ordered Weighted L1 Minimization

Thursday, December 14, 2017 - 4:30pm

Event Calendar Category

LIDS & Stats Tea

Speaker Name

Ulas Ayaz

Affiliation

LIDS

Building and Room Number

32-D677

Abstract

Ulas Ayaz will talk about geometry of subspace clustering following the work of Soltanolkotabi and Candes, and sparse estimation of strongly correlated variables using ordered weighted L1 regularizer following Figueiredo and Nowak. Additionally, he will present his ongoing work where he combines these two ideas to effectively improve subspace clustering by also using random geometric graphs.

Biography

Ulas Ayaz is postdoctoral associate at LIDS. He is working with Prof. Sertac Karaman on various robotic vision problems using compressed sensing. Prior to MIT, Ayaz was a postdoctoral fellow for a year at ICERM, Brown University. He has a PhD in Mathematics from the University of Bonn, Germany.