Uncertainty Quantification for Molecular Property Predictions with Graph Neural Architecture Search

TL;DR

AutoGNNUQ introduces an automated GNN architecture search method that effectively quantifies uncertainties in molecular property predictions, outperforming existing techniques and providing insights for dataset enhancement.

Contribution

The paper presents AutoGNNUQ, a novel automated approach combining architecture search and variance decomposition for uncertainty quantification in GNN-based molecular predictions.

Findings

AutoGNNUQ outperforms existing UQ methods in accuracy and uncertainty estimation.

The approach effectively separates aleatoric and epistemic uncertainties.

Visualization reveals correlations between molecular features and uncertainties.

Abstract

Graph Neural Networks (GNNs) have emerged as a prominent class of data-driven methods for molecular property prediction. However, a key limitation of typical GNN models is their inability to quantify uncertainties in the predictions. This capability is crucial for ensuring the trustworthy use and deployment of models in downstream tasks. To that end, we introduce AutoGNNUQ, an automated uncertainty quantification (UQ) approach for molecular property prediction. AutoGNNUQ leverages architecture search to generate an ensemble of high-performing GNNs, enabling the estimation of predictive uncertainties. Our approach employs variance decomposition to separate data (aleatoric) and model (epistemic) uncertainties, providing valuable insights for reducing them. In our computational experiments, we demonstrate that AutoGNNUQ outperforms existing UQ methods in terms of both prediction accuracy and UQ performance on multiple benchmark datasets. Additionally, we utilize t-SNE visualization to explore correlations between molecular features and uncertainty, offering insight for dataset improvement. AutoGNNUQ has broad applicability in domains such as drug discovery and materials science, where accurate uncertainty quantification is crucial for decision-making.