Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Generalized Evidential Deep Learning: From a Bayesian Perspective

About

Evidential Deep Learning (EDL) has emerged as an efficient, sampling-free strategy for uncertainty estimation. A series of EDL variants have been proposed to address specific limitations of the original framework, achieving notable success. However, the underlying theoretical structure of EDL and the relationships among these variants have received limited systematic investigation. In this work, we establish a principled theoretical foundation for EDL by interpreting it within a generalized Bayesian framework that includes prior specification, posterior update, and training objective. We further characterize evidential uncertainty from a Bayesian distributional uncertainty viewpoint, established via asymptotic analysis. Building on this perspective, we further propose Generalized Evidential Deep Learning (GEDL), a unified and extensible framework that explicitly disentangles the roles of individual components and systematically relates GEDL to existing variants. Extensive experiments demonstrate that GEDL yields comparable results on classification, uncertainty estimation and OOD detections, with theoretical grounding.

Yuanye Liu, Yibo Gao, Yuanyang Chen, Xiahai Zhuang• 2026

Related benchmarks

TaskDatasetResultRank
Model CalibrationCIFAR10 (test)
ECE4.64
68
Out-of-Distribution DetectionCIFAR-10 → SVHN--
14
Out-of-Distribution DetectionCIFAR-10 CIFAR-100--
14
OOD DetectionMNIST → KMNIST--
13
Image ClassificationMNIST
Accuracy99.58
9
Image ClassificationCIFAR-10
Accuracy90.22
9
Out-of-Distribution DetectionMNIST → KMNIST
Mean Prediction Probability (MP)99.8
9
Out-of-Distribution DetectionMNIST → FMNIST
Mean Performance (MP)98.54
9
ID ClassificationMNIST
Accuracy99.58
7
OOD DetectionCIFAR-10 vs CIFAR-100
Mean Precision (MP)88.27
7
Showing 10 of 13 rows

Other info

Follow for update