Timezone: »

 
Poster
Global Convergence of Least Squares EM for Demixing Two Log-Concave Densities
Wei Qian · Yuqian Zhang · Yudong Chen

Tue Dec 10 05:30 PM -- 07:30 PM (PST) @ East Exhibition Hall B + C #119
in Optimization -- Non-Convex Optimization »

This work studies the location estimation problem for a mixture of two rotation invariant log-concave densities. We demonstrate that Least Squares EM, a variant of the EM algorithm, converges to the true location parameter from a randomly initialized point. Moreover, we establish the explicit convergence rates and sample complexity bounds, revealing their dependence on the signal-to-noise ratio and the tail property of the log-concave distributions. Our analysis generalizes previous techniques for proving the convergence results of Gaussian mixtures, and highlights that an angle-decreasing property is sufficient for establishing global convergence for Least Squares EM.

Keywords: Learning Theory · Non-Convex Optimization · Optimization · Algorithms -> Density Estimation; Probabilistic Methods -> Latent Variable Models; Theory
[ Paper]  [ Poster] 

Author Information

Wei Qian (Cornell Univeristy)
Yuqian Zhang (Rutgers University)
Yudong Chen (Cornell University)

More from the Same Authors