When In Doubt, Abstain: The Impact of Abstention on Strategic Classification

2025-10-15

AI Summary

This paper investigates how classifier abstention mechanisms can mitigate strategic feature manipulation by agents in strategic classification. We address the problem wherein agents strategically alter observable features to obtain favorable decisions. To this end, we propose a Stackelberg game-theoretic framework in which the decision-maker proactively employs abstention as a defensive strategy against manipulation. We theoretically establish that the optimal abstention policy strictly improves the decision-maker’s utility without sacrificing accuracy—outperforming the no-abstention baseline. Crucially, when manipulation costs are low, abstention significantly raises the manipulation threshold for low-qualification agents, effectively deterring unnecessary manipulation. This work is the first to systematically demonstrate that abstention functions not merely as an error-tolerance mechanism but also as an economic incentive instrument to counter strategic behavior. Our findings introduce a novel paradigm for designing robust, incentive-aware machine learning algorithms.

Technology Category

• Multiagent Systems: Mechanism Design
• Game Theory and Economic Paradigms: Adversarial Learning
• Machine Learning: Feature Construction/Reformulation

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Algorithmic decision making is increasingly prevalent, but often vulnerable to strategic manipulation by agents seeking a favorable outcome. Prior research has shown that classifier abstention (allowing a classifier to decline making a decision due to insufficient confidence) can significantly increase classifier accuracy. This paper studies abstention within a strategic classification context, exploring how its introduction impacts strategic agents' responses and how principals should optimally leverage it. We model this interaction as a Stackelberg game where a principal, acting as the classifier, first announces its decision policy, and then strategic agents, acting as followers, manipulate their features to receive a desired outcome. Here, we focus on binary classifiers where agents manipulate observable features rather than their true features, and show that optimal abstention ensures that the principal's utility (or loss) is no worse than in a non-abstention setting, even in the presence of strategic agents. We also show that beyond improving accuracy, abstention can also serve as a deterrent to manipulation, making it costlier for agents, especially those less qualified, to manipulate to achieve a positive outcome when manipulation costs are significant enough to affect agent behavior. These results highlight abstention as a valuable tool for reducing the negative effects of strategic behavior in algorithmic decision making systems.

Problem

• Studying abstention's impact on strategic classification with Stackelberg game modeling
• Ensuring principal's utility is not worsened by strategic agents through abstention
• Using abstention as a manipulation deterrent by increasing agent manipulation costs

Innovation

• Introduces abstention in strategic classification games
• Models interaction as Stackelberg game with agents
• Shows abstention deters manipulation by increasing costs