Courtroom Analogy: New Perspective on Uncertainty-Aware Classification
2026-05-25 • Machine Learning
Machine Learning
AI summaryⓘ
The authors propose a new way to measure uncertainty in classification tasks by imagining a courtroom debate where each class has an advocate giving an opinion. These opinions are represented using a mathematical tool called Dirichlet distributions, which are combined based on how plausible each advocate is given the input. Their method, called MoDEX, uses a neural network to predict these parameters in one go, making the uncertainty measure both efficient and easier to understand. Experiments show that their approach not only works well but also provides clearer insights into how uncertainty is formed and combined.
Uncertainty QuantificationClassificationDirichlet DistributionNeural NetworksMixture ModelsInterpretabilitySingle-pass InferenceProbabilistic Modeling
Authors
Taeseong Yoon, Heeyoung Kim
Abstract
Single-pass uncertainty quantification (UQ) methods for classification represent uncertainty by predicting a tractable distribution over the class probability vector. While existing approaches primarily focus on enhancing the expressiveness of this distribution, they often provide limited insight into how predictive uncertainty is structured and aggregated, resulting in weak interpretability. We introduce the courtroom analogy, which conceptualizes uncertainty-aware classification as a structured debate among class-specific advocates. Each advocate forms a probabilistic opinion, and a final verdict is reached by aggregating these opinions using input-dependent plausibility weights. In this framework, each advocate's opinion is modeled as a Dirichlet distribution whose concentration parameter is decomposed into shared evidence and class-specific advocacy. This yields a structured mixture of Dirichlet distributions with semantically interpretable parameters. To instantiate this formulation, we propose Mixture of Dirichlet EXperts (MoDEX), a single-pass neural architecture that predicts the courtroom parameters, enabling efficient and expressive UQ while explicitly modeling uncertainty aggregation. We demonstrate that MoDEX enjoys strong theoretical properties and achieves state-of-the-art UQ performance across diverse benchmarks, yielding interpretable uncertainty estimates with meaningful semantics.