Timezone: »

 
Poster
Smoothness Matrices Beat Smoothness Constants: Better Communication Compression Techniques for Distributed Optimization
Mher Safaryan · Filip Hanzely · Peter Richtarik

Fri Dec 10 08:30 AM -- 10:00 AM (PST) @
in Poster Session 8 »

Large scale distributed optimization has become the default tool for the training of supervised machine learning models with a large number of parameters and training data. Recent advancements in the field provide several mechanisms for speeding up the training, including {\em compressed communication}, {\em variance reduction} and {\em acceleration}. However, none of these methods is capable of exploiting the inherently rich data-dependent smoothness structure of the local losses beyond standard smoothness constants. In this paper, we argue that when training supervised models, {\em smoothness matrices}---information-rich generalizations of the ubiquitous smoothness constants---can and should be exploited for further dramatic gains, both in theory and practice. In order to further alleviate the communication burden inherent in distributed optimization, we propose a novel communication sparsification strategy that can take full advantage of the smoothness matrices associated with local losses. To showcase the power of this tool, we describe how our sparsification technique can be adapted to three distributed optimization algorithms---DCGD, DIANA and ADIANA---yielding significant savings in terms of communication complexity. The new methods always outperform the baselines, often dramatically so.

Keywords: Theory · Optimization · Machine Learning

Author Information

Mher Safaryan (KAUST)
Filip Hanzely (University of Edinburgh)

I am a PhD student of Optimization and Operational research at the University of Edinburgh under the supervision of Peter Richtárik, working on stochastic non-convex optimization.

Peter Richtarik (KAUST)

More from the Same Authors