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

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 adiverseset of final candidates that capture many optimal and near-optimal design configurations. We proposeDiversityInAdversarialModel-basedOptimization (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 adistribution matching problemwhere 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.