Predictive Multiplicity in Classification

By Charles Marx, Flávio du Pin Calmon, Berk Ustun. 2019

210 citations recorded by the source corpus.

Prediction problems often admit competing models that perform almost equally well. This effect challenges key assumptions in machine learning when competing models assign conflicting predictions. In this paper, we define predictive multiplicity as the ability of a prediction problem to admit competing models with conflicting predictions. We introduce formal measures to evaluate the severity of predictive multiplicity and develop integer programming tools to compute them exactly for linear classification problems. We apply our tools to measure predictive multiplicity in recidivism prediction problems. Our results show that real-world datasets may admit competing models that assign wildly conflicting predictions, and motivate the need to measure and report predictive multiplicity in model development.

Subfields: Artificial Intelligence, Computational Theory and Mathematics, Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computational Theory and Mathematics

In Safety, Robustness & Interpretability

Citations in the corpus

Underspecification Presents Challenges for Credibility in Modern Machine Learning (2020, background): “Our treatment of underspecification is more closely related to work on “Rashomon sets” (Fisher et al., 2019), “predictive multiplicity” (Marx et al., 2019), and methods that search for validationequivalent predictors that are “right for the right reasons” (Ross et al., 2017).”
Interpretable Machine Learning: Fundamental Principles and 10 Grand Challenges (2021, methodology): “The Rashomon set includes these multiple descriptions of the same dataset (and again, none of them are assumed to be the truth, see Semenova et al., 2019; Fisher et al., 2019; Dong and Rudin, 2020; Marx et al., 2020).”
Characterizing Fairness Over the Set of Good Models Under Selective Labels (2021, background): “” Even though they have similar accuracy, these models may differ along other key dimensions, and recent work considers the implications of the Rashomon effect for model simplicity, interpretability, and explainability (Fisher et al., 2019; Marx et al., 2020; Rudin, 2019; Dong & Rudin, 2020; Semenov”
Model Multiplicity: Opportunities, Concerns, and Solutions (2022, background): “While some dimensions of exploration may be costly, such as collecting more data to explore alternate features to include in a model, the most commonmethod of exploration—hyperparameter variation [40, 51]—is already standard machine learning practice.”
It's COMPASlicated: The Messy Relationship between RAI Datasets and Algorithmic Fairness Benchmarks (2021, background): “Given the idea of predictive multiplicity– that multiple different predictive models achieve similar aggregate accuracy [54] – different data processing decisions may better realize normative objectives like not penalizing nowdecriminalized behavior or not allowing one’s past mistakes to haunt them ”