Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
Generalized Regularized Evidential Deep Learning Models: Theory and Comprehensive Evaluation
Deep Shankar Pandey, Hyomin Choi, Qi Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add correct evidence regularization (GRED) | Exp evidential model · CIFAR-100 | accuracy (%): 75.76 -> 76.00 | Improved |
| add correct evidence regularization (GRED) | SELU evidential model · MNIST | accuracy (%): 98.08 -> 98.68 | Improved |
| add correct evidence regularization (GRED) | SoftPlus evidential model · Tiny-ImageNet | accuracy (%): 90.25 -> 90.63 | Improved |
| add correct evidence regularization (GRED) | Exp evidential model · CIFAR-10 | accuracy (%): 95.24 -> 95.44 | Improved |
| use ReLU instead of smooth evidential activations (ablation) | evidential classifier · CIFAR-10 | accuracy (%): 95.15 -> 19.50 | Hurt |
| add correct evidence regularization (GRED) | Exp evidential CodeFormer · CelebA blind face restoration | PSNR: 21.46 -> 21.79 | Improved |
| add correct evidence regularization (GRED) | SELU evidential CodeFormer · CelebA blind face restoration | MSE: 514.82 -> 451.93 | Improved |
| add uncertainty-guided top-t belief-weighted codebook selection (t=10) | GRED-Exp CodeFormer · CelebA blind face restoration | PSNR: 21.79 -> 22.33 | Improved |
| increase top-t from 5 to 10 (ablation) | GRED-Exp CodeFormer · CelebA blind face restoration | MSE: 409.64 -> 403.69 | Improved |
| add correct evidence regularization (GRED) | Exp evidential VPT · 100-way 5-shot CIFAR-100 | accuracy (%): 49.05 -> 77.00 | Improved |
| add correct evidence regularization (GRED) | Exp evidential VPT · 100-way 1-shot CIFAR-100 | accuracy (%): 37.49 -> 46.54 | Improved |
| use evidential MSE loss instead of log or Type-II losses (ablation) | Exp evidential model · MNIST | test accuracy | Hurt |
Closest papers by meaning
- Uncertainty Regularized Evidential Regression (AAAI 2024 · similarity 0.91)
- Are Uncertainty Quantification Capabilities of Evidential Deep Learning a Mirage? (NeurIPS 2024 · similarity 0.90)
- The Evidence Contraction Issue in Deep Evidential Regression: Discussion and Solution (AAAI 2024 · similarity 0.90)
- Uncertainty Estimation by Flexible Evidential Deep Learning (NeurIPS 2025 · similarity 0.89)
- Calibrating LLMs with Information-Theoretic Evidential Deep Learning (ICLR 2025 · similarity 0.88)
- ANEDL: Adaptive Negative Evidential Deep Learning for Open-Set Semi-supervised Learning (AAAI 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
