LoCoDL: Communication-Efficient Distributed Learning with Local Training and Compression
This work addresses communication bottlenecks in distributed and federated learning, which is critical for scalable machine learning systems, though it is incremental as it builds on established techniques like local training and compression.
The paper tackles the problem of high communication costs in distributed and federated learning by introducing LoCoDL, an algorithm that combines local training and compression to reduce communication frequency and data size, achieving a doubly-accelerated communication complexity in heterogeneous settings with strongly convex functions and outperforming existing methods in practice.
In Distributed optimization and Learning, and even more in the modern framework of federated learning, communication, which is slow and costly, is critical. We introduce LoCoDL, a communication-efficient algorithm that leverages the two popular and effective techniques of Local training, which reduces the communication frequency, and Compression, in which short bitstreams are sent instead of full-dimensional vectors of floats. LoCoDL works with a large class of unbiased compressors that includes widely-used sparsification and quantization methods. LoCoDL provably benefits from local training and compression and enjoys a doubly-accelerated communication complexity, with respect to the condition number of the functions and the model dimension, in the general heterogenous regime with strongly convex functions. This is confirmed in practice, with LoCoDL outperforming existing algorithms.