Sample-Efficient Optimization in the Latent Space of Deep Generative Models via Weighted Retraining

TL;DR

This paper presents a sample-efficient black-box optimization method that leverages a deep generative model's latent space, actively retraining and weighting it to improve optimization performance in high-dimensional, structured input spaces.

Contribution

It introduces a novel weighted retraining approach that enhances the latent space for more efficient optimization, outperforming existing methods.

Findings

Significantly improves optimization efficiency on synthetic problems.

Enhances performance on real-world drug design tasks.

Easily integrates with existing generative model-based optimization methods.

Abstract

Many important problems in science and engineering, such as drug design, involve optimizing an expensive black-box objective function over a complex, high-dimensional, and structured input space. Although machine learning techniques have shown promise in solving such problems, existing approaches substantially lack sample efficiency. We introduce an improved method for efficient black-box optimization, which performs the optimization in the low-dimensional, continuous latent manifold learned by a deep generative model. In contrast to previous approaches, we actively steer the generative model to maintain a latent manifold that is highly useful for efficiently optimizing the objective. We achieve this by periodically retraining the generative model on the data points queried along the optimization trajectory, as well as weighting those data points according to their objective function value. This weighted retraining can be easily implemented on top of existing methods, and is empirically shown to significantly improve their efficiency and performance on synthetic and real-world optimization problems.