Bayesian Graph Neural Networks for Molecular Property Prediction
arXiv:2012.02089
Abstract
Graph neural networks for molecular property prediction are frequently underspecified by data and fail to generalise to new scaffolds at test time. A potential solution is Bayesian learning, which can capture our uncertainty in the model parameters. This study benchmarks a set of Bayesian methods applied to a directed MPNN, using the QM9 regression dataset. We find that capturing uncertainty in both readout and message passing parameters yields enhanced predictive accuracy, calibration, and performance on a downstream molecular search task.
arXiv commentPresented at NeurIPS 2020 Machine Learning for Molecules workshop
References in corpus (12)
- On Calibration of Modern Neural Networks
- Weight Uncertainty in Neural Networks
- Stochastic Gradient Hamiltonian Monte Carlo
- Variational Dropout and the Local Reparameterization Trick
- Averaging Weights Leads to Wider Optima and Better Generalization
- A Simple Baseline for Bayesian Uncertainty in Deep Learning
- Cyclical Stochastic Gradient MCMC for Bayesian Deep Learning
- Parallel and Distributed Thompson Sampling for Large-scale Accelerated Exploration of Chemical Space
- The Case for Bayesian Deep Learning
- Meta-Learning GNN Initializations for Low-Resource Molecular Property Prediction
- Subspace Inference for Bayesian Deep Learning
- A benchmark study on reliable molecular supervised learning via Bayesian learning