## Accelerating SGD for Highly Ill-Conditioned Huge-Scale Online Matrix Completion

### Jialun Zhang · Hong-Ming Chiu · Richard Y Zhang

##### Hall J #835

Keywords: [ SGD ] [ Stochastic Gradient Descent ] [ preconditioned gradient descent ] [ scaled gradient descent ] [ PrecGD ] [ huge scale optimization ] [ ScaledGD ]

[ Abstract ]
[ [
Thu 1 Dec 9 a.m. PST — 11 a.m. PST

Abstract: The matrix completion problem seeks to recover a $d\times d$ ground truth matrix of low rank $r\ll d$ from observations of its individual elements. Real-world matrix completion is often a huge-scale optimization problem, with $d$ so large that even the simplest full-dimension vector operations with $O(d)$ time complexity become prohibitively expensive. Stochastic gradient descent (SGD) is one of the few algorithms capable of solving matrix completion on a huge scale, and can also naturally handle streaming data over an evolving ground truth. Unfortunately, SGD experiences a dramatic slow-down when the underlying ground truth is ill-conditioned; it requires at least $O(\kappa\log(1/\epsilon))$ iterations to get $\epsilon$-close to ground truth matrix with condition number $\kappa$. In this paper, we propose a preconditioned version of SGD that preserves all the favorable practical qualities of SGD for huge-scale online optimization while also making it agnostic to $\kappa$. For a symmetric ground truth and the Root Mean Square Error (RMSE) loss, we prove that the preconditioned SGD converges to $\epsilon$-accuracy in $O(\log(1/\epsilon))$ iterations, with a rapid linear convergence rate as if the ground truth were perfectly conditioned with $\kappa=1$. In our numerical experiments, we observe a similar acceleration forill-conditioned matrix completion under the root mean square error (RMSE) loss, Euclidean distance matrix (EDM) completion under pairwise square loss, and collaborative filtering under the Bayesian Personalized Ranking (BPR) loss.

Chat is not available.