Pretrained Optimization Model for Zero-Shot Black Box Optimization

TL;DR

This paper introduces a Pretrained Optimization Model (POM) that effectively addresses zero-shot black-box optimization by leveraging prior knowledge, outperforming existing methods especially in high-dimensional and diverse tasks.

Contribution

The paper presents a novel pretrained optimization model that generalizes across tasks, dimensions, and horizons, reducing the need for hyperparameter tuning and improving zero-shot optimization performance.

Findings

POM outperforms state-of-the-art methods on BBOB and robot control tasks.

Fine-tuning POM with few samples significantly enhances performance.

POM demonstrates robust generalization across various task settings.

Abstract

Zero-shot optimization involves optimizing a target task that was not seen during training, aiming to provide the optimal solution without or with minimal adjustments to the optimizer. It is crucial to ensure reliable and robust performance in various applications. Current optimizers often struggle with zero-shot optimization and require intricate hyperparameter tuning to adapt to new tasks. To address this, we propose a Pretrained Optimization Model (POM) that leverages knowledge gained from optimizing diverse tasks, offering efficient solutions to zero-shot optimization through direct application or fine-tuning with few-shot samples. Evaluation on the BBOB benchmark and two robot control tasks demonstrates that POM outperforms state-of-the-art black-box optimization methods, especially for high-dimensional tasks. Fine-tuning POM with a small number of samples and budget yields significant performance improvements. Moreover, POM demonstrates robust generalization across diverse task distributions, dimensions, population sizes, and optimization horizons. For code implementation, see https://github.com/ninja-wm/POM/.