Performative Risk Control: Calibrating Models for Reliable Deployment under Performativity

2025-05-30

AI Summary

This work addresses the black-box model calibration problem under *performative prediction*—where model predictions influence the true outcomes. We propose the first risk-controlling calibration framework with finite-sample statistical guarantees. Methodologically, we integrate rigorous statistical risk control into the performative learning paradigm, designing an iterative refinement calibration algorithm whose risk convergence and predictive performance improvement are theoretically established. The framework supports multiple risk measures and Hoeffding-type tail bounds. Empirically, on credit default prediction, it significantly improves risk coverage and calibration accuracy while maintaining robustness and controllability under distribution shift. Our approach thus provides both verifiable theoretical guarantees and practical tools for trustworthy decision-making in dynamic environments.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Calibrating blackbox machine learning models to achieve risk control is crucial to ensure reliable decision-making. A rich line of literature has been studying how to calibrate a model so that its predictions satisfy explicit finite-sample statistical guarantees under a fixed, static, and unknown data-generating distribution. However, prediction-supported decisions may influence the outcome they aim to predict, a phenomenon named performativity of predictions, which is commonly seen in social science and economics. In this paper, we introduce Performative Risk Control, a framework to calibrate models to achieve risk control under performativity with provable theoretical guarantees. Specifically, we provide an iteratively refined calibration process, where we ensure the predictions are improved and risk-controlled throughout the process. We also study different types of risk measures and choices of tail bounds. Lastly, we demonstrate the effectiveness of our framework by numerical experiments on the task of predicting credit default risk. To the best of our knowledge, this work is the first one to study statistically rigorous risk control under performativity, which will serve as an important safeguard against a wide range of strategic manipulation in decision-making processes.

Problem

• Calibrating models for reliable risk control under performativity
• Ensuring predictions improve and remain risk-controlled iteratively
• Addressing strategic manipulation in decision-making with statistical guarantees

Innovation

• Iterative calibration for risk control
• Theoretical guarantees under performativity
• Multiple risk measures and bounds