Distributed Principal Component Analysis with Limited Communication

Part of Advances in Neural Information Processing Systems 34 (NeurIPS 2021)

Bibtex Paper Reviews And Public Comment » Supplemental

Authors

Foivos Alimisis, Peter Davies, Bart Vandereycken, Dan Alistarh

Abstract

We study efficient distributed algorithms for the fundamental problem of principal component analysis and leading eigenvector computation on the sphere, when the data are randomly distributed among a set of computational nodes. We propose a new quantized variant of Riemannian gradient descent to solve this problem, and prove that the algorithm converges with high probability under a set of necessary spherical-convexity properties. We give bounds on the number of bits transmitted by the algorithm under common initialization schemes, and investigate the dependency on the problem dimension in each case.