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

Energy-based Epistemic Uncertainty for Graph Neural Networks

About

In domains with interdependent data, such as graphs, quantifying the epistemic uncertainty of a Graph Neural Network (GNN) is challenging as uncertainty can arise at different structural scales. Existing techniques neglect this issue or only distinguish between structure-aware and structure-agnostic uncertainty without combining them into a single measure. We propose GEBM, an energy-based model (EBM) that provides high-quality uncertainty estimates by aggregating energy at different structural levels that naturally arise from graph diffusion. In contrast to logit-based EBMs, we provably induce an integrable density in the data space by regularizing the energy function. We introduce an evidential interpretation of our EBM that significantly improves the predictive robustness of the GNN. Our framework is a simple and effective post hoc method applicable to any pre-trained GNN that is sensitive to various distribution shifts. It consistently achieves the best separation of in-distribution and out-of-distribution data on 6 out of 7 anomaly types while having the best average rank over shifts on \emph{all} datasets.

Dominik Fuchsgruber, Tom Wollschl\"ager, Stephan G\"unnemann• 2024

Related benchmarks

TaskDatasetResultRank
Out-of-Distribution DetectionChameleon
Rank1.8
24
Out-of-Distribution DetectionPubmed
Rank1.8
24
Out-of-Distribution DetectionCoraML
Rank1.8
24
Out-of-Distribution Detectionamazon-ratings
Rank2.8
24
Out-of-Distribution DetectionSquirrel
Rank3
24
Out-of-Distribution DetectionRoman-Empire
Rank5.2
24
O.O.D. detectionAmazon Ratings Far-Features
AUC-PR (Alea.)30.8
12
O.O.D. detectionRoman Empire Far-Features
AUC-PR (Aleatoric)34.6
12
O.O.D. detectionRoman Empire (LoC)
AUC-PR (Aleatoric)36.9
12
Out-of-Distribution DetectionRoman Empire Local Class (LoC)
AUC-ROC (Aleatoric)73.3
12
Showing 10 of 17 rows

Other info

Follow for update