Variance-based Gradient Compression for Efficient Distributed Deep Learning
This addresses the problem of efficient distributed training for deep learning practitioners, particularly in low-bandwidth environments, though it appears incremental as it builds on existing gradient compression techniques.
The paper tackles the communication bottleneck in distributed deep learning by proposing a variance-based gradient compression method that delays updates until gradients are unambiguous, achieving high compression ratios while maintaining model accuracy.
Due to the substantial computational cost, training state-of-the-art deep neural networks for large-scale datasets often requires distributed training using multiple computation workers. However, by nature, workers need to frequently communicate gradients, causing severe bottlenecks, especially on lower bandwidth connections. A few methods have been proposed to compress gradient for efficient communication, but they either suffer a low compression ratio or significantly harm the resulting model accuracy, particularly when applied to convolutional neural networks. To address these issues, we propose a method to reduce the communication overhead of distributed deep learning. Our key observation is that gradient updates can be delayed until an unambiguous (high amplitude, low variance) gradient has been calculated. We also present an efficient algorithm to compute the variance with negligible additional cost. We experimentally show that our method can achieve very high compression ratio while maintaining the result model accuracy. We also analyze the efficiency using computation and communication cost models and provide the evidence that this method enables distributed deep learning for many scenarios with commodity environments.