Learning Credal Ensembles via Distributionally Robust Optimization

TL;DR

This paper introduces CreDRO, a distributionally robust ensemble method that captures epistemic uncertainty from distribution shifts, improving out-of-distribution detection and selective classification performance.

Contribution

It proposes a novel credal ensemble learning approach using distributionally robust optimization to better quantify epistemic uncertainty from distribution shifts.

Findings

Outperforms existing credal methods on OOD detection benchmarks

Improves selective classification in medical applications

Captures meaningful epistemic uncertainty beyond training randomness

Abstract

Credal predictors are models that are aware of epistemic uncertainty and produce a convex set of probabilistic predictions. They offer a principled way to quantify predictive epistemic uncertainty (EU) and have been shown to improve model robustness in various settings. However, most state-of-the-art methods mainly define EU as disagreement caused by random training initializations, which mostly reflects sensitivity to optimization randomness rather than uncertainty from deeper sources. To address this, we define EU as disagreement among models trained with varying relaxations of the i.i.d. assumption between training and test data. Based on this idea, we propose CreDRO, which learns an ensemble of plausible models through distributionally robust optimization. As a result, CreDRO captures EU not only from training randomness but also from meaningful disagreement due to potential distribution shifts between training and test data. Empirical results show that CreDRO consistently outperforms existing credal methods on tasks such as out-of-distribution detection across multiple benchmarks and selective classification in medical applications.