用扩散模型直接生成多目标优化解，无需额外建模。
Pareto-Conditioned Diffusion Models for Offline Multi-Objective Optimization
- 以帕累托权衡为条件，直接生成目标解。
- 在多个基准上表现优异且任务间一致性更强。
- 适合需要稳定泛化的离线多目标优化场景。
多目标优化（MOO）在现实应用中广泛存在，需平衡相互冲突的目标。在离线设置下，仅能使用静态数据集，主要挑战在于超越已观测数据的泛化能力。本文提出帕累托条件扩散模型（Pareto-Conditioned Diffusion, PCD），将离线MOO建模为条件采样问题。通过直接以期望的权衡关系为条件，PCD避免了显式代理模型的需求。为有效探索帕累托前沿，PCD采用重加权策略聚焦高绩效样本，并引入参考方向机制引导采样至训练数据之外的新颖、有前景区域。在标准离线MOO基准上的实验表明，PCD实现了极具竞争力的性能，且相比现有方法在多种任务间展现出更强的一致性。
原文摘要 · Abstract (English)
Multi-objective optimization (MOO) arises in many real-world applications where trade-offs between competing objectives must be carefully balanced. In the offline setting, where only a static dataset is available, the main challenge is generalizing beyond observed data. We introduce Pareto-Conditioned Diffusion (PCD), a novel framework that formulates offline MOO as a conditional sampling problem. By conditioning directly on desired trade-offs, PCD avoids the need for explicit surrogate models. To effectively explore the Pareto front, PCD employs a reweighting strategy that focuses on high-performing samples and a reference-direction mechanism to guide sampling towards novel, promising regions beyond the training data. Experiments on standard offline MOO benchmarks show that PCD achieves highly competitive performance and, importantly, demonstrates greater consistency across diverse tasks than existing offline MOO approaches.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。