Preference-Guided Diffusion for Multi-Objective Offline Optimization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Annadani, Yashas, Belakaria, Syrine, Ermon, Stefano, Bauer, Stefan, Engelhardt, Barbara E
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908719063760896
author Annadani, Yashas
Belakaria, Syrine
Ermon, Stefano
Bauer, Stefan
Engelhardt, Barbara E
author_facet Annadani, Yashas
Belakaria, Syrine
Ermon, Stefano
Bauer, Stefan
Engelhardt, Barbara E
contents Offline multi-objective optimization aims to identify Pareto-optimal solutions given a dataset of designs and their objective values. In this work, we propose a preference-guided diffusion model that generates Pareto-optimal designs by leveraging a classifier-based guidance mechanism. Our guidance classifier is a preference model trained to predict the probability that one design dominates another, directing the diffusion model toward optimal regions of the design space. Crucially, this preference model generalizes beyond the training distribution, enabling the discovery of Pareto-optimal solutions outside the observed dataset. We introduce a novel diversity-aware preference guidance, augmenting Pareto dominance preference with diversity criteria. This ensures that generated solutions are optimal and well-distributed across the objective space, a capability absent in prior generative methods for offline multi-objective optimization. We evaluate our approach on various continuous offline multi-objective optimization tasks and find that it consistently outperforms other inverse/generative approaches while remaining competitive with forward/ surrogate-based optimization methods. Our results highlight the effectiveness of classifier-guided diffusion models in generating diverse and high-quality solutions that approximate the Pareto front well.
format Preprint
id arxiv_https___arxiv_org_abs_2503_17299
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Preference-Guided Diffusion for Multi-Objective Offline Optimization
Annadani, Yashas
Belakaria, Syrine
Ermon, Stefano
Bauer, Stefan
Engelhardt, Barbara E
Machine Learning
Artificial Intelligence
Offline multi-objective optimization aims to identify Pareto-optimal solutions given a dataset of designs and their objective values. In this work, we propose a preference-guided diffusion model that generates Pareto-optimal designs by leveraging a classifier-based guidance mechanism. Our guidance classifier is a preference model trained to predict the probability that one design dominates another, directing the diffusion model toward optimal regions of the design space. Crucially, this preference model generalizes beyond the training distribution, enabling the discovery of Pareto-optimal solutions outside the observed dataset. We introduce a novel diversity-aware preference guidance, augmenting Pareto dominance preference with diversity criteria. This ensures that generated solutions are optimal and well-distributed across the objective space, a capability absent in prior generative methods for offline multi-objective optimization. We evaluate our approach on various continuous offline multi-objective optimization tasks and find that it consistently outperforms other inverse/generative approaches while remaining competitive with forward/ surrogate-based optimization methods. Our results highlight the effectiveness of classifier-guided diffusion models in generating diverse and high-quality solutions that approximate the Pareto front well.
title Preference-Guided Diffusion for Multi-Objective Offline Optimization
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2503.17299