arXiv · Machine Learning Theory· Philipp J. Schneider, Lukas Looser, Antoine Garin, Shuhan Liu, Daniel Kuhn·· 3 小时前AI 评分25
WRAP:非平稳市场下深度对冲的对抗训练
Adversarial Training for Deep Hedging in Nonstationary Markets
AI 导读
论文提出 WRAP(Wasserstein-Reweighting Adversarial Perturbation),一种源自双预算分布鲁棒优化(DRO)的漂移感知对抗训练框架,用于改进深度对冲策略。
来源:arXiv · Machine Learning Theory · arxiv.org