Structure-aware graph neural network based deep transfer learning framework for enhanced predictive analytics on diverse materials datasets

Abstract

Modern data mining methods have demonstrated effectiveness in comprehending and predicting materials properties. An essential component in the process of materials discovery is to know which material(s) will possess desirable properties. For many materials properties, performing experiments and density functional theory computations are costly and time-consuming. Hence, it is challenging to build accurate predictive models for such properties using conventional data mining methods due to the small amount of available data. Here we present a framework for materials property prediction tasks using structure information that leverages graph neural network-based architecture along with deep-transfer-learning techniques to drastically improve the model’s predictive ability on diverse materials (3D/2D, inorganic/organic, computational/experimental) data. We evaluated the proposed framework in cross-property and cross-materials class scenarios using 115 datasets to find that transfer learning models outperform the models trained from scratch in 104 cases, i.e., ≈90%, with additional benefits in performance for extrapolation problems. We believe the proposed framework can be widely useful in accelerating materials discovery in materials science.

From bulk effective mass to 2D carrier mobility accurate prediction via adversarial transfer learning

“…Nevertheless, the effectiveness of this approach for systems with higher degrees of dissimilarity, such as from ordered crystal to disordered materials like alloy, remains untested and may not be as successful. It may require an improved adversarial training approaches or borrowing generative adversarial network methods from inverse design 40 – 43 . Another significant challenge in this field is how to select the most appropriate source-domain tasks from various options available.…”