Performative Prediction on Games and Mechanism Design

2024-08-09

arXiv.org

AI Summary

This paper investigates the feedback effects of “executive predictions” in collective-risk dilemmas—where rational agents dynamically adjust strategies based on predictions, thereby altering the very collective outcomes being predicted. Prior work neglects strategic interdependence between predictors and predicted agents. Method: We formalize the dynamic feedback of predictions on multi-agent behavior via a Bayesian trust-updating model and propose a mechanism-design framework that jointly optimizes prediction accuracy and social welfare under rationality constraints. Contribution/Results: We identify a paradox wherein higher prediction accuracy degrades aggregate welfare. Theoretically, we prove that unconditionally prioritizing prediction stability likely harms social welfare. Empirically, our mechanism achieves Pareto-improving trade-offs between predictive fidelity and welfare, significantly enhancing equilibrium social welfare over baseline approaches.

Technology Category

• Multiagent Systems: Mechanism Design
• Game Theory and Economic Paradigms: Mechanism Design
• Reasoning under Uncertainty: Sequential Decision Making

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Agents often have individual goals which depend on a group's actions. If agents trust a forecast of collective action and adapt strategically, such prediction can influence outcomes non-trivially, resulting in a form of performative prediction. This effect is ubiquitous in scenarios ranging from pandemic predictions to election polls, but existing work has ignored interdependencies among predicted agents. As a first step in this direction, we study a collective risk dilemma where agents dynamically decide whether to trust predictions based on past accuracy. As predictions shape collective outcomes, social welfare arises naturally as a metric of concern. We explore the resulting interplay between accuracy and welfare, and demonstrate that searching for stable accurate predictions can minimize social welfare with high probability in our setting. By assuming knowledge of a Bayesian agent behavior model, we then show how to achieve better trade-offs and use them for mechanism design.

Problem

• Study performative prediction in collective risk dilemmas
• Explore interplay between prediction accuracy and social welfare
• Develop mechanism design for better prediction-welfare trade-offs

Innovation

• Performative prediction in games
• Bayesian agent behavior model
• Mechanism design for welfare optimization