/ames_interpretability

Uncertainty aware and explainable Bayesian neural network for prediction of Ames mutagenicity

Primary LanguageJupyter NotebookMIT LicenseMIT

Uncertainty aware and explainable Bayesian neural network for prediction of Ames mutagenicity

Official repository for the paper "Model explanation methods can reproduce structural alerts in Bayesian neural networks for risk assessment of mutagenicity"

Author: Charles Gong

Bayesian neural network for prediction of Ames mutagenicity. Contains methods for uncertainty quantification and model explanations, as well as experiments and validation.