Quasi-global Momentum: Accelerating Decentralized Deep Learning on Heterogeneous Data

Tao Lin · Sai Praneeth Reddy Karimireddy · Sebastian Stich · Martin Jaggi

Keywords: [ Optimization for Deep Networks ]

[ Abstract ]
[ Paper ]
[ Visit Poster at Spot B0 in Virtual World ]
Wed 21 Jul 9 p.m. PDT — 11 p.m. PDT
Spotlight presentation: Deep Learning Optimization
Wed 21 Jul 5 p.m. PDT — 6 p.m. PDT


Decentralized training of deep learning models is a key element for enabling data privacy and on-device learning over networks. In realistic learning scenarios, the presence of heterogeneity across different clients' local datasets poses an optimization challenge and may severely deteriorate the generalization performance. In this paper, we investigate and identify the limitation of several decentralized optimization algorithms for different degrees of data heterogeneity. We propose a novel momentum-based method to mitigate this decentralized training difficulty. We show in extensive empirical experiments on various CV/NLP datasets (CIFAR-10, ImageNet, and AG News) and several network topologies (Ring and Social Network) that our method is much more robust to the heterogeneity of clients' data than other existing methods, by a significant improvement in test performance (1%-20%).

Chat is not available.