Amortized Implicit Differentiation for Stochastic Bilevel Optimization

Author name not available (Why is that?)

Publication date: 29 November 2021

Abstract: We study a class of algorithms for solving bilevel optimization problems in both stochastic and deterministic settings when the inner-level objective is strongly convex. Specifically, we consider algorithms based on inexact implicit differentiation and we exploit a warm-start strategy to amortize the estimation of the exact gradient. We then introduce a unified theoretical framework inspired by the study of singularly perturbed systems (Habets, 1974) to analyze such amortized algorithms. By using this framework, our analysis shows these algorithms to match the computational complexity of oracle methods that have access to an unbiased estimate of the gradient, thus outperforming many existing results for bilevel optimization. We illustrate these findings on synthetic experiments and demonstrate the efficiency of these algorithms on hyper-parameter optimization experiments involving several thousands of variables.

Has companion code repository: https://github.com/michaelarbel/amigo

This page was built for publication: Amortized Implicit Differentiation for Stochastic Bilevel Optimization

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6384225)