Uncertainty Modeling in Graph Neural Networks via Stochastic Differential Equations

Abstract

We address the problem of learning uncertainty-aware representations for graph-structured data. While Graph Neural Ordinary Differential Equations (GNODE) are effective in learning node representations, they fail to quantify uncertainty. To address this, we introduce Latent Graph Neural Stochastic Differential Equations (LGNSDE), which enhance GNODE by embedding randomness through Brownian motion to quantify uncertainty. We provide theoretical guarantees for LGNSDE and empirically show better performance in uncertainty quantification.

PDF NeurIPSW OpenReview Semantic Scholar

Cite

Text

Bergna et al. "Uncertainty Modeling in Graph Neural Networks via Stochastic Differential Equations." NeurIPS 2024 Workshops: BDU, 2024.

Markdown

[Bergna et al. "Uncertainty Modeling in Graph Neural Networks via Stochastic Differential Equations." NeurIPS 2024 Workshops: BDU, 2024.](https://mlanthology.org/neuripsw/2024/bergna2024neuripsw-uncertainty/)

BibTeX

@inproceedings{bergna2024neuripsw-uncertainty,
  title     = {{Uncertainty Modeling in Graph Neural Networks via Stochastic Differential Equations}},
  author    = {Bergna, Richard and Ordoñez, Sergio Calvo and Opolka, Felix and Lio, Pietro and Hernández-Lobato, José Miguel},
  booktitle = {NeurIPS 2024 Workshops: BDU},
  year      = {2024},
  url       = {https://mlanthology.org/neuripsw/2024/bergna2024neuripsw-uncertainty/}
}