AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Uncertainty Estimation by Fisher Information-based Evidential Deep Learning

Danruo DENG, Guangyong Chen, Yang YU, Furui Liu, Pheng-Ann Heng

ICML 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 12 citations (OpenAlex)

Read the paper

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 changedWhereMeasuredEffect
Use Fisher-information-weighted evidential learningEDL on CIFAR10 classification · CIFAR10accuracy: 83.55 ± 0.6 -> 89.20 ± 0.3Improved
Use Fisher-information-weighted evidential learningEDL on OOD detection · MNIST → KMNISTAUPR (Max.P): 97.02 ± 0.8 -> 98.34 ± 0.2Improved
Use Fisher-information-weighted evidential learningEDL on few-shot classification · mini-ImageNet 10-way 20-shotaccuracy: 69.29 ± 0.17 -> 78.60 ± 0.08Improved
Add the negative log-determinant term |I| only (ablation)EDL · mini-ImageNet 5-way 5-shotaccuracy: 80.38 ± 0.15 -> 81.82 ± 0.14Improved
Add the negative log-determinant term |I| only (ablation)EDL · mini-ImageNet 5-way 5-shotAUPR (OOD, α₀): 76.53 ± 0.27 -> 79.68 ± 0.25Improved
Add the negative log-determinant term |I| only (ablation)EDL · mini-ImageNet 5-way 5-shotAUPR (confidence, Max.α): 93.92 ± 0.09 -> 93.97 ± 0.09No clear effect
Replace MSE with FIM-weighted MSE only (ablation)EDL · mini-ImageNet 5-way 5-shotaccuracy: 80.38 ± 0.15 -> 81.27 ± 0.14Improved
Replace MSE with FIM-weighted MSE only (ablation)EDL · mini-ImageNet 5-way 5-shotAUPR (confidence, Max.α): 93.92 ± 0.09 -> 94.42 ± 0.08Improved
Replace MSE with FIM-weighted MSE only (ablation)EDL · mini-ImageNet 5-way 5-shotAUPR (OOD, α₀): 76.53 ± 0.27 -> 81.75 ± 0.22Improved
Use both FIM-weighted MSE and the |I| term (ablation)EDL · mini-ImageNet 5-way 5-shotaccuracy: 80.38 ± 0.15 -> 82.00 ± 0.14Improved
Use both FIM-weighted MSE and the |I| term (ablation)EDL · mini-ImageNet 5-way 5-shotAUPR (confidence, Max.α): 93.92 ± 0.09 -> 94.09 ± 0.09Improved
Use both FIM-weighted MSE and the |I| term (ablation)EDL · mini-ImageNet 5-way 5-shotAUPR (OOD, α₀): 76.53 ± 0.27 -> 82.48 ± 0.20Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy