Vicinal Label Supervision for Reliable Aleatoric and Epistemic Uncertainty Estimation
Linye Li, Yufei Chen, Xiaodong Yue
Abstract
Uncertainty estimation is crucial for ensuring the reliability of machine learning models in safety-critical applications. Evidential Deep Learning (EDL) offers a principled framework by modeling predictive uncertainty through Dirichlet distributions over class probabilities. However, existing EDL methods predominantly rely on level-0 hard labels, which supervise an uncertainty-aware model with full certainty. We argue that hard labels not only fail to capture epistemic uncertainty but also obscure the aleatoric uncertainty arising from inherent data noise and label ambiguity. As a result, EDL models often produce degenerate Dirichlet distributions that collapse to near-deterministic outputs. To overcome these limitations, we propose a vicinal risk minimization paradigm for EDL by incorporating level-1 supervision in the form of vicinally smoothed conditional label distributions. This richer supervision exposes the model to local label uncertainty, enhancing aleatoric uncertainty quantification while mitigating the degeneration of the Dirichlet distribution into a Dirac delta function, thereby improving epistemic uncertainty modeling. Extensive experiments show that our approach consistently outperforms standard EDL baselines across synthetic datasets, covariate-shifted out-of-distribution generalization tasks, out-of-distribution detection, and selective classification benchmarks, providing more reliable uncertainty estimates.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
Cited by top-tier papers1
Builds on26
- Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance AwarenessJeremiah Z. Liu, Zi Lin, Shreyas Padhy, Dustin Tran et al.NeurIPS 2020 · 604 citations
- Uncertainty Estimation Using a Single Deep Deterministic Neural NetworkJoost van Amersfoort, Lewis Smith, Yee Whye Teh, Yarin GalICML 2020 · 529 citations
- Human Uncertainty Makes Classification More RobustJoshua C. Peterson, Ruairidh M. Battleday, Thomas L. Griffiths, Olga RussakovskyICCV 2019 · 362 citations
- Posterior Network: Uncertainty Estimation without OOD Samples via Density-Based Pseudo-CountsBertrand Charpentier, Daniel Zügner, Stephan GünnemannNeurIPS 2020 · 263 citations
- Reliable Conflictive Multi-View LearningCai Xu, Jiajun Si, Ziyu Guan, Wei Zhao et al.AAAI 2024 · 121 citations
Related papers
- Uncertainty Estimation by Flexible Evidential Deep LearningTaeseong Yoon, Heeyoung KimNeurIPS 2025 · 12 citations
- Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage?Maohao Shen, Jongha Jon Ryu, Soumya Ghosh, Yuheng Bu et al.NeurIPS 2024 · 29 citations
- Uncertainty Estimation by Density Aware Evidential Deep LearningTaeseong Yoon, Heeyoung KimICML 2024 · 16 citations
- Is Epistemic Uncertainty Faithfully Represented by Evidential Deep Learning Methods?Mira Jürgens, Nis Meinert, Viktor Bengs, Eyke Hüllermeier et al.ICML 2024 · 35 citations
- R-EDL: Relaxing Nonessential Settings of Evidential Deep LearningMengyuan Chen, Junyu Gao, Changsheng XuICLR 2024 · 18 citations
