Averaging Rate Scheduler for Decentralized Learning on Heterogeneous Data
This work addresses data heterogeneity in decentralized learning, offering an incremental improvement for practical scenarios where data distributions are non-IID.
The paper tackles the problem of data heterogeneity in decentralized learning by proposing an averaging rate scheduler, which improves test accuracy by approximately 3% compared to constant averaging rates.
State-of-the-art decentralized learning algorithms typically require the data distribution to be Independent and Identically Distributed (IID). However, in practical scenarios, the data distribution across the agents can have significant heterogeneity. In this work, we propose averaging rate scheduling as a simple yet effective way to reduce the impact of heterogeneity in decentralized learning. Our experiments illustrate the superiority of the proposed method (~3% improvement in test accuracy) compared to the conventional approach of employing a constant averaging rate.