Learnable optimization algorithms improve themselves through code evolution
Hyper Algorithm Design Agent: Evolving Learnable Optimizer from Zero
Artificial Intelligence
Summary
Optimization methods help computers find the best solutions to problems, but designing these methods usually requires human experts and careful tweaking. The paper presents a system where one software agent writes and improves optimization algorithms, while a second agent improves the first agent itself. This creates a cycle where the system keeps evolving better algorithms automatically, learning to adapt efficiently to different kinds of optimization challenges. Their experiments show this approach outperforms existing human-designed algorithms and adapts quickly to new problem types.
What this means in practice
- •For machine learning engineers: Create optimization algorithms that self-improve and adapt efficiently to different machine learning tasks without manual redesign.
- •For robotics developers: Develop adaptable optimization routines that evolve automatically for robot control and decision-making tasks in changing environments.
Authors
Zipei Yu, Yue-Jiao Gong, Zeyuan Ma, Yuncheng Jiang, Zhiguang Cao
Abstract
Meta-Black-Box Optimization (MetaBBO) is one of the highlights in the recent AI for Optimization trend. This paradigm's bi-level workflow leverages the learnable algorithm design policy at meta level to ensure the performance and generalization improvement on the low-level optimization task. While MetaBBO helps advance the performance lower bound of the resulted optimization system, it is currently handcrafted and customized case by case to adapt different optimization problems, which inevitably introduces inherent subjectivity and hence restricts the performance upper bound and usability in practice. In this paper, we address this issue by regarding MetaBBO's design loop as coding task, where we could introduce openendedness into MetaBBO with recursive self-improvement capability of advanced coding agents. Specifically, we propose a dual-agent framework: i) a task agent continuously refines the codebase of a target MetaBBO approach through code evolution; ii) a hyper agent progressively modifies the task agent and itself to provide open-ended design behavior; iii) the evolved MetaBBO codebase is evaluated and all in-execution information is fed back to the agents for recursive self-referential improvement. As a result, given a naive MetaBBO template, our framework automates a design evolution and finds novel variants superior to up-to-date human-made MetaBBO baselines. Surprisingly, the experimental results also demonstrate that our framework supports fast adaption across different optimization domains. Solid interpretation analysis further reveals interesting design principles emerge in such open-ended process. This work serves as the first exploration on automating design of complex learning-assisted optimization algorithms.