Who's Gaming the System? A Causally-Motivated Approach for Detecting Strategic Adaptation

TL;DR

This paper introduces a causally-motivated framework to detect and rank agents who manipulate their inputs to machine learning models, addressing the challenge of identifying the most aggressive gaming agents without knowing their utility functions.

Contribution

It formulates gaming detection as a causal effect estimation problem, enabling the ranking of agents by their gaming propensity even with partial identifiability.

Findings

Validated approach with synthetic data demonstrating effective detection.

Case study in diagnosis coding reveals features linked to gaming behavior.

Framework successfully ranks agents by gaming intensity.

Abstract

In many settings, machine learning models may be used to inform decisions that impact individuals or entities who interact with the model. Such entities, or agents, may game model decisions by manipulating their inputs to the model to obtain better outcomes and maximize some utility. We consider a multi-agent setting where the goal is to identify the "worst offenders:" agents that are gaming most aggressively. However, identifying such agents is difficult without knowledge of their utility function. Thus, we introduce a framework in which each agent's tendency to game is parameterized via a scalar. We show that this gaming parameter is only partially identifiable. By recasting the problem as a causal effect estimation problem where different agents represent different "treatments," we prove that a ranking of all agents by their gaming parameters is identifiable. We present empirical results in a synthetic data study validating the usage of causal effect estimation for gaming detection and show in a case study of diagnosis coding behavior in the U.S. that our approach highlights features associated with gaming.