Title	(Quasi)-Efficiently Learning Mixtures of Gaussians at the Statistically Optimal Separation
Creator	Kothari, Pravesh
Publisher	Banff International Research Station for Mathematical Innovation and Discovery
Date Issued	2018-08-15T11:07
Description	Recovering a hidden signal or structure in the presence of random noise is a recurring theme in fundamental problems arising in computational complexity, cryptography, machine learning, and statistics. In the recent years, a Sum-of-Squares method, a hierarchy of generic semi-definite programming relaxations, has yielded a systematic approach for such "parameter estimation" problems. In this talk, I'll illustrate the SoS method for parameter estimation by means of a recent application of to learning mixture of gaussians with information theoretically optimal cluster-separation in quasi-polynomial time. No sub-exponential time algorithm was previously known in this regime. Based on joint works with Jacob Steinhardt and David Steurer.
Extent	61.0
Subject	Mathematics; Computer science; Fourier analysis; Theoretical computer science
Type	Moving Image
File Format	video/mp4
Language	eng
Notes	Author affiliation: Princeton University and Institute for Advanced Study
Series	BIRS Workshop Lecture Videos (Oaxaca de Juárez (Mexico))
Date Available	2019-03-29
Provider	Vancouver : University of British Columbia Library
Rights	Attribution-NonCommercial-NoDerivatives 4.0 International
DOI	10.14288/1.0377647
URI	http://hdl.handle.net/2429/69329
Affiliation	Non UBC
Peer Review Status	Unreviewed
Scholarly Level	Postdoctoral
Rights URI	http://creativecommons.org/licenses/by-nc-nd/4.0/
Aggregated Source Repository	DSpace

Open Collections

BIRS Workshop Lecture Videos

(Quasi)-Efficiently Learning Mixtures of Gaussians at the Statistically Optimal Separation Kothari, Pravesh

Description

Item Metadata

Item Media

Item Citations and Data

Rights