INTERACT: Achieving Low Sample and Communication Complexities in Decentralized Bilevel Learning over Networks

TL;DR

This paper introduces INTERACT and SVR-INTERACT algorithms that achieve low sample and communication complexities for decentralized bilevel optimization over networks, addressing key challenges in distributed machine learning.

Contribution

The paper presents the first algorithms with provably low sample and communication complexities for decentralized bilevel optimization with nonconvex and strongly-convex structures.

Findings

INTERACT requires O(n ε^{-1}) sample complexity and O(ε^{-1}) communication complexity.

SVR-INTERACT reduces sample complexity to O(√n ε^{-1}) while maintaining communication efficiency.

Numerical experiments validate the theoretical advantages of the proposed algorithms.

Abstract

In recent years, decentralized bilevel optimization problems have received increasing attention in the networking and machine learning communities thanks to their versatility in modeling decentralized learning problems over peer-to-peer networks (e.g., multi-agent meta-learning, multi-agent reinforcement learning, personalized training, and Byzantine-resilient learning). However, for decentralized bilevel optimization over peer-to-peer networks with limited computation and communication capabilities, how to achieve low sample and communication complexities are two fundamental challenges that remain under-explored so far. In this paper, we make the first attempt to investigate the class of decentralized bilevel optimization problems with nonconvex and strongly-convex structure corresponding to the outer and inner subproblems, respectively. Our main contributions in this paper are two-fold: i) We first propose a deterministic algorithm called INTERACT (inner-gradient-descent-outer-tracked-gradient) that requires the sample complexity of $\mathcal{O}(n \epsilon^{-1})$ and communication complexity of $\mathcal{O}(\epsilon^{-1})$ to solve the bilevel optimization problem, where $n$ and $\epsilon > 0$ are the number of samples at each agent and the desired stationarity gap, respectively. ii) To relax the need for full gradient evaluations in each iteration, we propose a stochastic variance-reduced version of INTERACT (SVR-INTERACT), which improves the sample complexity to $\mathcal{O}(\sqrt{n} \epsilon^{-1})$ while achieving the same communication complexity as the deterministic algorithm. To our knowledge, this work is the first that achieves both low sample and communication complexities for solving decentralized bilevel optimization problems over networks. Our numerical experiments also corroborate our theoretical findings.