Rashomon Capacity: A Metric for Predictive Multiplicity in Classification

TL;DR

The paper introduces Rashomon Capacity, a new metric for measuring predictive multiplicity in probabilistic classifiers, helping to identify conflicting predictions among models with similar performance.

Contribution

It proposes Rashomon Capacity as a novel, rigorous metric for predictive multiplicity applicable to probabilistic classifiers, with practical estimation methods and implications for model transparency.

Findings

Rashomon Capacity effectively captures predictive multiplicity across datasets.

The metric provides a principled way to disclose conflicting models to stakeholders.

Experiments show its applicability to neural networks and various models.

Abstract

Predictive multiplicity occurs when classification models with statistically indistinguishable performances assign conflicting predictions to individual samples. When used for decision-making in applications of consequence (e.g., lending, education, criminal justice), models developed without regard for predictive multiplicity may result in unjustified and arbitrary decisions for specific individuals. We introduce a new metric, called Rashomon Capacity, to measure predictive multiplicity in probabilistic classification. Prior metrics for predictive multiplicity focus on classifiers that output thresholded (i.e., 0-1) predicted classes. In contrast, Rashomon Capacity applies to probabilistic classifiers, capturing more nuanced score variations for individual samples. We provide a rigorous derivation for Rashomon Capacity, argue its intuitive appeal, and demonstrate how to estimate it in practice. We show that Rashomon Capacity yields principled strategies for disclosing conflicting models to stakeholders. Our numerical experiments illustrate how Rashomon Capacity captures predictive multiplicity in various datasets and learning models, including neural networks. The tools introduced in this paper can help data scientists measure and report predictive multiplicity prior to model deployment.