Cross-Gradient Aggregation for Decentralized Learning from Non-IID data
Date
2021-06-28
Authors
Esfandiari, Yasaman
Tan, Sin Yong
Jiang, Zhanhong
Balu, Aditya
Herron, Ethan
Major Professor
Advisor
Committee Member
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Decentralized learning enables a group of collaborative agents to learn models using a distributed dataset without the need for a central parameter server. Recently, decentralized learning algorithms have demonstrated state-of-the-art results on benchmark data sets, comparable with centralized algorithms. However, the key assumption to achieve competitive performance is that the data is independently and identically distributed (IID) among the agents which, in real-life applications, is often not applicable. Inspired by ideas from continual learning, we propose Cross-Gradient Aggregation (CGA), a novel decentralized learning algorithm where (i) each agent aggregates cross-gradient information, i.e., derivatives of its model with respect to its neighbors' datasets, and (ii) updates its model using a projected gradient based on quadratic programming (QP). We theoretically analyze the convergence characteristics of CGA and demonstrate its efficiency on non-IID data distributions sampled from the MNIST and CIFAR-10 datasets. Our empirical comparisons show superior learning performance of CGA over existing state-of-the-art decentralized learning algorithms, as well as maintaining the improved performance under information compression to reduce peer-to-peer communication overhead. The code is available here on GitHub.
Series Number
Journal Issue
Is Version Of
Versions
Series
Academic or Administrative Unit
Type
Preprint
Comments
This is a preprint from Esfandiari, Yasaman, Sin Yong Tan, Zhanhong Jiang, Aditya Balu, Ethan Herron, Chinmay Hegde, and Soumik Sarkar. "Cross-Gradient Aggregation for Decentralized Learning from Non-IID data." arXiv e-prints (2021): arXiv-2103. doi: https://doi.org/10.48550/arXiv.2103.02051.