Mistake, Manipulation and Margin Guarantees in Online Strategic Classification

arXiv:2403.18176v1 Announce Type: new
Abstract: We consider an online strategic classification problem where each arriving agent can manipulate their true feature vector to obtain a positive predicted label, while incurring a cost that depends on the amount of manipulation. The learner seeks to predict the agent's true label given access to only the manipulated features. After the learner releases their prediction, the agent's true label is revealed. Previous algorithms such as the strategic perceptron guarantee finitely many mistakes under a …

abstract agent arxiv classification cost cs.gt cs.lg feature feature vector manipulation math.oc positive true type vector