A Loopless Distributed Algorithm for Personalized Bilevel Optimization

Youcheng Niu,Ying Sun,Yan Huang,Jinming Xu
DOI: https://doi.org/10.1109/CDC49753.2023.10383649
2023-01-01
Abstract:This paper studies a class of personalized distributed bilevel optimization problems over networks, where nodes aim at jointly optimizing the sum of outer-level objectives that depend on the solution of inner-level optimization problems. The existing algorithms for distributed bilevel optimization problems usually require extra computation loops for estimating hypergradients. To facilitate computational efficiency, we develop a loopless distributed algorithm that employs certain steps to approximate the optimal solution of inner-level optimization problems, and track Hessian-inverse-vector products in a recursive manner. We prove that for stochastic nonconvex-strongly-convex problems, the proposed algorithm achieves the state of the art O(epsilon(-2)) communication cost, while improving the computational cost by O(log(1/epsilon)). Numerical experiments validate our theoretical findings.
What problem does this paper attempt to address?