Credal and Interval Deep Evidential Classifications

2025-12-05

AI Summary

This work addresses the challenge of jointly modeling epistemic (reducible) and aleatoric (irreducible) uncertainty in classification tasks. We propose two novel deep learning methods—CDEC and IDEC—that are the first to integrate credal sets and evidential intervals into deep classification frameworks, enabling fine-grained uncertainty quantification and automatic prediction abstention. Our approach formulates an evidence-theoretic loss function, allowing standard backpropagation to train neural networks that output evidential distributions. Crucially, reliable uncertainty estimation is achieved even with small ensembles. Extensive experiments on MNIST, CIFAR, and multiple out-of-distribution (OOD) benchmarks demonstrate that our models retain competitive classification accuracy while significantly improving OOD detection performance and predictive calibration. This yields enhanced robustness and reliability in the presence of distributional shift.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Computer Vision: Learning & Optimization for CV

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Uncertainty Quantification (UQ) presents a pivotal challenge in the field of Artificial Intelligence (AI), profoundly impacting decision-making, risk assessment and model reliability. In this paper, we introduce Credal and Interval Deep Evidential Classifications (CDEC and IDEC, respectively) as novel approaches to address UQ in classification tasks. CDEC and IDEC leverage a credal set (closed and convex set of probabilities) and an interval of evidential predictive distributions, respectively, allowing us to avoid overfitting to the training data and to systematically assess both epistemic (reducible) and aleatoric (irreducible) uncertainties. When those surpass acceptable thresholds, CDEC and IDEC have the capability to abstain from classification and flag an excess of epistemic or aleatoric uncertainty, as relevant. Conversely, within acceptable uncertainty bounds, CDEC and IDEC provide a collection of labels with robust probabilistic guarantees. CDEC and IDEC are trained using standard backpropagation and a loss function that draws from the theory of evidence. They overcome the shortcomings of previous efforts, and extend the current evidential deep learning literature. Through extensive experiments on MNIST, CIFAR-10 and CIFAR-100, together with their natural OoD shifts (F-MNIST/K-MNIST, SVHN/Intel, TinyImageNet), we show that CDEC and IDEC achieve competitive predictive accuracy, state-of-the-art OoD detection under epistemic and total uncertainty, and tight, well-calibrated prediction regions that expand reliably under distribution shift. An ablation over ensemble size further demonstrates that CDEC attains stable uncertainty estimates with only a small ensemble.

Problem

• Address uncertainty quantification in AI classification tasks
• Avoid overfitting and assess epistemic and aleatoric uncertainties
• Provide robust probabilistic guarantees and detect out-of-distribution shifts

Innovation

• Credal sets and evidential intervals for uncertainty quantification
• Abstention mechanism triggered by excessive epistemic or aleatoric uncertainty
• Training with evidence theory loss via standard backpropagation