生成式分布鲁棒优化
Generative Distributionally Robust Optimization
查看机构详情
- Telfer School of Management, University of Ottawa（渥太华大学泰尔弗管理学院）
- Department of Electrical and Computer Engineering, Western University（西安大略大学电气与计算机工程系）
机构由 AI 辅助整理，请以论文原文为准。
浏览论文内容
中文总结 AI 辅助
研究提出生成式分布鲁棒优化（GDRO）框架，接受可采样条件生成器为名义模型，通过采样器-辛科恩配对限制最坏情况法则，可直接有限样本近似和可微实现，相比名义决策降低了罕见上下文库存遗憾和SocialGAN导航碰撞。
中文摘要 AI 辅助
生成模型在分布鲁棒优化（DRO）中越来越多地被采用，但现有方法在模型兼容性和对抗结构之间进行权衡：接受任意采样器的方法不会将最坏情况法则限制在生成器族中，而基于生成器参数化的对手则依赖于特定于模型的访问，如似然、分数或训练数据。我们提出了生成式分布鲁棒优化（GDRO），这是一个有原则的框架，它接受任何可采样的条件生成器作为名义模型，并将最坏情况法则限制在选定的条件生成器族中。关键是采样器-辛科恩配对：采样器精确表示条件法则，而辛科恩散度在无需似然访问的情况下比较它们的诱导分布，并且可以仅从样本中估计。由此产生的总体问题允许在有效决策上下文中进行直接的有限样本近似和可微的原始对偶实现。对于利普希茨损失，总体辛科恩半径限制了下游退化。相对于名义决策，我们的方法在显式和隐式生成器中，将罕见上下文库存遗憾降低了60%，并将SocialGAN导航碰撞降低了50%。
英文摘要
Generative models are increasingly adopted in distributionally robust optimization (DRO), but existing approaches trade off model compatibility and adversarial structure: methods that accept arbitrary samplers do not restrict worst-case laws to a generator family, while generator-parameterized adversaries rely on model-specific access such as likelihoods, scores, or training data. We propose Generative Distributionally Robust Optimization (GDRO), a principled framework that accepts any sampleable conditional generator as the nominal model and restricts worst-case laws to a chosen conditional generator family. The key is the sampler-Sinkhorn pairing: samplers represent the conditional laws exactly, while Sinkhorn divergence compares their induced distributions without likelihood access and can be estimated from samples alone. The resulting population problem admits a direct finite-sample approximation and differentiable primal-dual implementation at the active decision context. For Lipschitz losses, the population Sinkhorn radius bounds downstream degradation. Across explicit and implicit generators, our method reduces rare-context inventory regret by 60% and SocialGAN navigation collisions by 50% relative to nominal decisions.
