Strategic Classification from Revealed Preferences

TL;DR

This paper introduces an online learning algorithm for linear classification in strategic settings where agents manipulate features, achieving low regret despite unknown utility functions and strategic behavior.

Contribution

It presents a computationally efficient method for learning classifiers in strategic environments with revealed preferences, ensuring near-optimal performance over time.

Findings

Algorithm achieves diminishing Stackelberg regret

Learner effectively responds to strategic agent manipulation

Method works for a broad family of agent cost functions

Abstract

We study an online linear classification problem, in which the data is generated by strategic agents who manipulate their features in an effort to change the classification outcome. In rounds, the learner deploys a classifier, and an adversarially chosen agent arrives, possibly manipulating her features to optimally respond to the learner. The learner has no knowledge of the agents' utility functions or "real" features, which may vary widely across agents. Instead, the learner is only able to observe their "revealed preferences" --- i.e. the actual manipulated feature vectors they provide. For a broad family of agent cost functions, we give a computationally efficient learning algorithm that is able to obtain diminishing "Stackelberg regret" --- a form of policy regret that guarantees that the learner is obtaining loss nearly as small as that of the best classifier in hindsight, even allowing for the fact that agents will best-respond differently to the optimal classifier.