Wasserstein Distributionally Robust Online Learning

arXiv:2602.20403v1 Announce Type: new Abstract: We study distributionally robust online learning, where a risk-averse learner updates decisions sequentially to guard against worst-case distributions drawn from …