End-to-end AI framework for interpretable prediction of molecular and crystal properties

TL;DR

This paper presents an integrated AI framework that combines hyperparameter optimization, accelerated training, and interpretability for predicting properties of molecules and crystals, demonstrated on multiple datasets and deployed on supercomputers.

Contribution

The authors develop a comprehensive, open-source AI framework that unifies state-of-the-art models, optimization, and interpretability for materials property prediction in high-performance computing environments.

Findings

Framework successfully predicts material properties on benchmark datasets.

Demonstrates transferability across molecules, crystals, and frameworks.

Achieves accelerated training and interpretability in supercomputing environments.

Abstract

We introduce an end-to-end computational framework that allows for hyperparameter optimization using the DeepHyper library, accelerated model training, and interpretable AI inference. The framework is based on state-of-the-art AI models including CGCNN, PhysNet, SchNet, MPNN, MPNN-transformer, and TorchMD-NET. We employ these AI models along with the benchmark QM9, hMOF, and MD17 datasets to showcase how the models can predict user-specified material properties within modern computing environments. We demonstrate transferable applications in the modeling of small molecules, inorganic crystals and nanoporous metal organic frameworks with a unified, standalone framework. We have deployed and tested this framework in the ThetaGPU supercomputer at the Argonne Leadership Computing Facility, and in the Delta supercomputer at the National Center for Supercomputing Applications to provide researchers with modern tools to conduct accelerated AI-driven discovery in leadership-class computing environments. We release these digital assets as open source scientific software in GitLab, and ready-to-use Jupyter notebooks in Google Colab.