A Provably Convergent Plug-and-Play Framework for Stochastic Bilevel Optimization
This work addresses a foundational problem in machine learning for researchers and practitioners dealing with hierarchical optimization tasks, offering a theoretically sound and flexible framework.
The paper tackles the challenge of developing efficient stochastic bilevel optimization methods by proposing a plug-and-play framework (PnPBO) that integrates various stochastic estimators, achieving optimal sample complexity comparable to single-level optimization and resolving an open question in the field.
Bilevel optimization has recently attracted significant attention in machine learning due to its wide range of applications and advanced hierarchical optimization capabilities. In this paper, we propose a plug-and-play framework, named PnPBO, for developing and analyzing stochastic bilevel optimization methods. This framework integrates both modern unbiased and biased stochastic estimators into the single-loop bilevel optimization framework introduced in [9], with several improvements. In the implementation of PnPBO, all stochastic estimators for different variables can be independently incorporated, and an additional moving average technique is applied when using an unbiased estimator for the upper-level variable. In the theoretical analysis, we provide a unified convergence and complexity analysis for PnPBO, demonstrating that the adaptation of various stochastic estimators (including PAGE, ZeroSARAH, and mixed strategies) within the PnPBO framework achieves optimal sample complexity, comparable to that of single-level optimization. This resolves the open question of whether the optimal complexity bounds for solving bilevel optimization are identical to those for single-level optimization. Finally, we empirically validate our framework, demonstrating its effectiveness on several benchmark problems and confirming our theoretical findings.