Diversity By Design: Leveraging Distribution Matching for Offline Model-Based Optimization

Michael S Yao, James Gee, Osbert Bastani

Track: AI for Science

Keywords: Offline Optimization, AI4Science, Generative Design, Model-Based Optimization

TL;DR: We introduce a method to propose both high-quality and diverse designs in a wide variety of offline optimization tasks in AI4Science.

Abstract: The goal of offline model-based optimization (MBO) is to propose new designs that maximize a reward function given only an offline dataset. However, an important desiderata is to also propose a *diverse* set of final candidates that capture many optimal and near-optimal design configurations. We propose **D**iversit**y** i**n** **A**dversarial **M**odel-based **O**ptimization (**DynAMO**) as a novel method to introduce design diversity as an explicit objective into any MBO problem. Our key insight is to formulate diversity as a *distribution matching problem* where the distribution of generated designs captures the inherent diversity contained within the offline dataset. Extensive experiments spanning multiple scientific domains show that DynAMO can be used with common optimization methods to significantly improve the diversity of proposed designs while still discovering high-quality candidates.

Serve As Reviewer: ~Michael_S_Yao1, ~Osbert_Bastani1

Submission Number: 6