Pretrained model rapidly generates solutions for multi-objective optimization
In-Context Preference-Conditioned Pareto Set Learning from Limited Offline Data
Artificial Intelligence
Summary
Sometimes, when trying to balance different goals in a problem, people want to find the best trade-offs between them. The authors show a way to pretrain a model on lots of example problems so it can quickly create a tailored solution for a new problem using only existing data, without needing extra trial-and-error. This method saves time by producing solutions for any given preference in just one step. The authors tested their approach on both made-up and real problems and found it made good trade-off solutions faster than existing methods.
What this means in practice
- •For optimization engineers: Generate solutions balancing multiple goals quickly for new problems using only pre-existing data without retraining.
- •For machine learning developers: Use a pretrained hypernetwork to create problem-specific models that handle preference-conditioned optimization in a single forward pass.
Authors
Yiming Yao, Liang Zhao, Qingfu Zhang
Abstract
Multi-objective optimization involves diverse preferences over conflicting objectives. Rather than returning a finite set of trade-off solutions, preference-conditioned Pareto set learning seeks to map a specified preference directly to a corresponding Pareto optimal solution. Yet in many real-world scenarios, optimization must proceed offline without further objective evaluations, leaving only a limited static dataset that sparsely covers the decision and objective spaces. Existing methods either optimize separately for each preference or train a problem-specific Pareto solution model, requiring repeated search across preferences or repeated training across problems. We introduce in-context preference-conditioned Pareto set learning (ICPSL), a framework that amortizes both preference-wise solution search and problem-specific model construction. ICPSL pretrains a foundation-style hypernetwork on a large and diverse set of synthetic prior data. Given the offline observations of an unseen problem, the pretrained hypernetwork learns to generate a problem-specific, preference-conditioned Pareto solution model in a single forward pass. The generated model can then be reused to predict solutions for arbitrary preferences without problem-specific training or preference-wise search. Experiments on synthetic and real-world problems show that ICPSL produces high-quality Pareto solutions while providing substantial speedups in problem-specific solution-model construction.