Collaborative Deep Learning in Fixed Topology Networks
This work addresses the need for decentralized deep learning for agents with local data in constrained networks, but it appears incremental as it builds on existing distributed and federated learning methods.
The paper tackles the problem of parallelizing deep learning in decentralized, communication-constrained environments by proposing a consensus-based distributed SGD algorithm (CDSGD) that enables data parallelization without a central server. It demonstrates competitive performance compared to centralized SGD and federated averaging on datasets like MNIST, CIFAR-10, and CIFAR-100, though specific numerical gains are not detailed in the abstract.
There is significant recent interest to parallelize deep learning algorithms in order to handle the enormous growth in data and model sizes. While most advances focus on model parallelization and engaging multiple computing agents via using a central parameter server, aspect of data parallelization along with decentralized computation has not been explored sufficiently. In this context, this paper presents a new consensus-based distributed SGD (CDSGD) (and its momentum variant, CDMSGD) algorithm for collaborative deep learning over fixed topology networks that enables data parallelization as well as decentralized computation. Such a framework can be extremely useful for learning agents with access to only local/private data in a communication constrained environment. We analyze the convergence properties of the proposed algorithm with strongly convex and nonconvex objective functions with fixed and diminishing step sizes using concepts of Lyapunov function construction. We demonstrate the efficacy of our algorithms in comparison with the baseline centralized SGD and the recently proposed federated averaging algorithm (that also enables data parallelism) based on benchmark datasets such as MNIST, CIFAR-10 and CIFAR-100.