Graph Neural Stochastic Differential Equations

TL;DR

This paper introduces Graph Neural SDEs, a new model that incorporates stochasticity into graph neural networks to improve uncertainty estimation and out-of-distribution detection.

Contribution

It proposes a novel stochastic differential equation framework for graph neural networks, enhancing uncertainty quantification over existing deterministic models.

Findings

Latent Graph Neural SDEs outperform traditional GNNs and GNN ODEs in confidence prediction.

The model effectively detects out-of-distribution data in static and spatio-temporal settings.

Empirical results demonstrate superior performance in uncertainty estimation.

Abstract

We present a novel model Graph Neural Stochastic Differential Equations (Graph Neural SDEs). This technique enhances the Graph Neural Ordinary Differential Equations (Graph Neural ODEs) by embedding randomness into data representation using Brownian motion. This inclusion allows for the assessment of prediction uncertainty, a crucial aspect frequently missed in current models. In our framework, we spotlight the \textit{Latent Graph Neural SDE} variant, demonstrating its effectiveness. Through empirical studies, we find that Latent Graph Neural SDEs surpass conventional models like Graph Convolutional Networks and Graph Neural ODEs, especially in confidence prediction, making them superior in handling out-of-distribution detection across both static and spatio-temporal contexts.