Uncertainty Quantification for Molecular Property Predictions with Graph Neural Architecture Search

arXiv:2307.10438v2 Announce Type: replace
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 …

abstract architecture arxiv capability class cs.lg data data-driven gnn gnns graph graph neural networks however key networks neural architecture search neural networks physics.chem-ph prediction predictions property q-bio.bm quantification search trustworthy type uncertainty