The Decoupled Risk Landscape in Performative Prediction

2025-06-10

AI Summary

This paper addresses the core challenge in performative prediction—distributional shifts induced by model deployment. We propose a dual-parameter decoupled risk landscape visualization method, explicitly formulating the loss as a joint function of model parameters and data-generation parameters for the first time, thereby geometrically exposing the intrinsic relationship between stable points and optimal solutions. Furthermore, we define and formalize the “extended performative prediction” paradigm, modeling strategic agent responses under partial observability—i.e., when agents only observe the outputs of a black-box deployed model. Through two-stage optimization modeling, nonlinear strategic classification simulations, and empirical convergence diagnostics, we demonstrate that existing algorithms frequently fail under nonlinear strategic behavior. Our framework provides an interpretable risk diagnostic tool, establishing both theoretical foundations and visual analytics support for policy-robust learning.

Technology Category

• Game Theory and Economic Paradigms: Adversarial Learning
• Reasoning under Uncertainty: Stochastic Optimization
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• 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

Abstract

Performative Prediction addresses scenarios where deploying a model induces a distribution shift in the input data, such as individuals modifying their features and reapplying for a bank loan after rejection. Literature has had a theoretical perspective giving mathematical guarantees for convergence (either to the stable or optimal point). We believe that visualization of the loss landscape can complement this theoretical advances with practical insights. Therefore, (1) we introduce a simple decoupled risk visualization method inspired in the two-step process that performative prediction is. Our approach visualizes the risk landscape with respect to two parameter vectors: model parameters and data parameters. We use this method to propose new properties of the interest points, to examine how existing algorithms traverse the risk landscape and perform under more realistic conditions, including strategic classification with non-linear models. (2) Building on this decoupled risk visualization, we introduce a novel setting - extended Performative Prediction - which captures scenarios where the distribution reacts to a model different from the decision-making one, reflecting the reality that agents often lack full access to the deployed model.

Problem

• Visualizing loss landscape in performative prediction scenarios
• Analyzing risk landscape with model and data parameters
• Introducing extended performative prediction for realistic model reactions

Innovation

• Decoupled risk visualization for performative prediction
• Extended Performative Prediction for realistic scenarios
• Visualizing risk landscape with model and data parameters