Uncertainty Estimation by Density Aware Evidential Deep Learning

Taeseong Yoon, Heeyoung Kim
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:57217-57243, 2024.

Abstract

Evidential deep learning (EDL) has shown remarkable success in uncertainty estimation. However, there is still room for improvement, particularly in out-of-distribution (OOD) detection and classification tasks. The limited OOD detection performance of EDL arises from its inability to reflect the distance between the testing example and training data when quantifying uncertainty, while its limited classification performance stems from its parameterization of the concentration parameters. To address these limitations, we propose a novel method called Density Aware Evidential Deep Learning (DAEDL). DAEDL integrates the feature space density of the testing example with the output of EDL during the prediction stage, while using a novel parameterization that resolves the issues in the conventional parameterization. We prove that DAEDL enjoys a number of favorable theoretical properties. DAEDL demonstrates state-of-the-art performance across diverse downstream tasks related to uncertainty estimation and classification.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-yoon24a, title = {Uncertainty Estimation by Density Aware Evidential Deep Learning}, author = {Yoon, Taeseong and Kim, Heeyoung}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {57217--57243}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/yoon24a/yoon24a.pdf}, url = {https://proceedings.mlr.press/v235/yoon24a.html}, abstract = {Evidential deep learning (EDL) has shown remarkable success in uncertainty estimation. However, there is still room for improvement, particularly in out-of-distribution (OOD) detection and classification tasks. The limited OOD detection performance of EDL arises from its inability to reflect the distance between the testing example and training data when quantifying uncertainty, while its limited classification performance stems from its parameterization of the concentration parameters. To address these limitations, we propose a novel method called Density Aware Evidential Deep Learning (DAEDL). DAEDL integrates the feature space density of the testing example with the output of EDL during the prediction stage, while using a novel parameterization that resolves the issues in the conventional parameterization. We prove that DAEDL enjoys a number of favorable theoretical properties. DAEDL demonstrates state-of-the-art performance across diverse downstream tasks related to uncertainty estimation and classification.} }
Endnote
%0 Conference Paper %T Uncertainty Estimation by Density Aware Evidential Deep Learning %A Taeseong Yoon %A Heeyoung Kim %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-yoon24a %I PMLR %P 57217--57243 %U https://proceedings.mlr.press/v235/yoon24a.html %V 235 %X Evidential deep learning (EDL) has shown remarkable success in uncertainty estimation. However, there is still room for improvement, particularly in out-of-distribution (OOD) detection and classification tasks. The limited OOD detection performance of EDL arises from its inability to reflect the distance between the testing example and training data when quantifying uncertainty, while its limited classification performance stems from its parameterization of the concentration parameters. To address these limitations, we propose a novel method called Density Aware Evidential Deep Learning (DAEDL). DAEDL integrates the feature space density of the testing example with the output of EDL during the prediction stage, while using a novel parameterization that resolves the issues in the conventional parameterization. We prove that DAEDL enjoys a number of favorable theoretical properties. DAEDL demonstrates state-of-the-art performance across diverse downstream tasks related to uncertainty estimation and classification.
APA
Yoon, T. & Kim, H.. (2024). Uncertainty Estimation by Density Aware Evidential Deep Learning. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:57217-57243 Available from https://proceedings.mlr.press/v235/yoon24a.html.

Related Material