In this paper, we concentrate on decentralized optimization problems with nonconvex and nonsmooth objective functions, especially on the decentralized training of nonsmooth neural networks. We introduce a unified framework to analyze the global convergence of decentralized stochastic subgradient-based methods. We prove the global convergence of our proposed framework under mild conditions, by establishing that the generated sequence asymptotically approximates the trajectories of its associated differential inclusion. Furthermore, we establish that our proposed framework covers a wide range of existing efficient decentralized subgradient-based methods, including decentralized stochastic subgradient descent (DSGD), DSGD with gradient-tracking technique (DSGD-T), and DSGD with momentum (DSGD-M). In addition, we introduce the sign map to regularize the update directions in DSGD-M, and show it is enclosed in our proposed framework. Consequently, our convergence results establish, for the first time, global convergence of these methods when applied to nonsmooth nonconvex objectives. Preliminary numerical experiments demonstrate that our proposed framework yields highly efficient decentralized subgradient-based methods with convergence guarantees in the training of nonsmooth neural networks.
翻译:暂无翻译