Timezone: »

Towards More Efficient Stochastic Decentralized Learning: Faster Convergence and Sparse Communication
Zebang Shen · Aryan Mokhtari · Tengfei Zhou · Peilin Zhao · Hui Qian

Wed Jul 11 02:40 AM -- 02:50 AM (PDT) @ A9

Recently, the decentralized optimization problem is attracting growing attention. Most existing methods are deterministic with high per-iteration cost and have a convergence rate quadratically depending on the problem condition number. Besides, the dense communication is necessary to ensure the convergence even if the dataset is sparse. In this paper, we generalize the decentralized optimization problem to a monotone operator root finding problem, and propose a stochastic algorithm named DSBA that (1) converges geometrically with a rate linearly depending on the problem condition number, and (2) can be implemented using sparse communication only. Additionally, DSBA handles important learning problems like AUC-maximization which can not be tackled efficiently in the previous problem setting. Experiments on convex minimization and AUC-maximization validate the efficiency of our method.

Author Information

Zebang Shen (Zhejiang University)
Aryan Mokhtari (MIT)
Tengfei Zhou (Zhejiang University)
Peilin Zhao (Artificial Intelligence Department, Ant ​Financial)
Hui Qian (Zhejiang University)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors