Decoupling Corruption and Horizon in Robust Contextual Pricing
arXiv:2607.11210
The paper proposes an online algorithm for repeated contextual pricing that tolerates a bounded number of corrupted sale feedbacks, achieving regret that scales with the corruption budget and context dimension but not with the time horizon.
Abstract
We study robust repeated contextual pricing, where valuations depends linearly on the features. At each round , a seller observes a context, posts a price, and receives only a possibly corrupted binary sale feedback. The seller knows an upper bound on the number of corrupted rounds. We design an algorithm with regret , where is the context dimension. This is the first guarantee for robust contextual pricing that separates the dependence on the corruption budget from the horizon , closing the problem left open by Gupta, Guruganesh, Paes Leme, and Schneider (2025).