Robust Guided Diffusion for Offline Black-Box Optimization

TL;DR

The paper introduces RGD, a novel method combining proxy guidance and proxy-free diffusion to improve offline black-box optimization, achieving state-of-the-art results on design tasks.

Contribution

It proposes a new framework, RGD, that integrates proxy-enhanced sampling and diffusion-based proxy refinement for robust, guided diffusion in offline optimization.

Findings

RGD outperforms existing methods on multiple design-bench tasks.

The approach effectively balances robustness and explicit guidance.

State-of-the-art results demonstrate its practical effectiveness.

Abstract

Offline black-box optimization aims to maximize a black-box function using an offline dataset of designs and their measured properties. Two main approaches have emerged: the forward approach, which learns a mapping from input to its value, thereby acting as a proxy to guide optimization, and the inverse approach, which learns a mapping from value to input for conditional generation. (a) Although proxy-free~(classifier-free) diffusion shows promise in robustly modeling the inverse mapping, it lacks explicit guidance from proxies, essential for generating high-performance samples beyond the training distribution. Therefore, we propose \textit{proxy-enhanced sampling} which utilizes the explicit guidance from a trained proxy to bolster proxy-free diffusion with enhanced sampling control. (b) Yet, the trained proxy is susceptible to out-of-distribution issues. To address this, we devise the module \textit{diffusion-based proxy refinement}, which seamlessly integrates insights from proxy-free diffusion back into the proxy for refinement. To sum up, we propose \textit{\textbf{R}obust \textbf{G}uided \textbf{D}iffusion for Offline Black-box Optimization}~(\textbf{RGD}), combining the advantages of proxy~(explicit guidance) and proxy-free diffusion~(robustness) for effective conditional generation. RGD achieves state-of-the-art results on various design-bench tasks, underscoring its efficacy. Our code is at https://github.com/GGchen1997/RGD.