Quantifying Aleatoric and Epistemic Uncertainty in Machine Learning: Are Conditional Entropy and Mutual Information Appropriate Measures?
This work addresses foundational issues in uncertainty quantification for ML practitioners, potentially impacting model reliability and decision-making, but it is incremental as it critiques existing methods rather than proposing new ones.
The paper challenges the appropriateness of using conditional entropy and mutual information to quantify aleatoric and epistemic uncertainty in machine learning, identifying theoretical incoherencies and questioning additive decomposition, with experiments in computer vision tasks supporting these concerns.
The quantification of aleatoric and epistemic uncertainty in terms of conditional entropy and mutual information, respectively, has recently become quite common in machine learning. While the properties of these measures, which are rooted in information theory, seem appealing at first glance, we identify various incoherencies that call their appropriateness into question. In addition to the measures themselves, we critically discuss the idea of an additive decomposition of total uncertainty into its aleatoric and epistemic constituents. Experiments across different computer vision tasks support our theoretical findings and raise concerns about current practice in uncertainty quantification.