Improving neural network predictions of material properties with limited data using transfer learning
This work addresses the challenge of data-efficient modeling in materials science, enabling generalizable neural network models without costly large datasets, though it is incremental as it applies existing transfer learning concepts to a new domain.
The authors tackled the problem of predicting material properties from ab initio simulations with limited data by developing new transfer learning algorithms, achieving improved predictions for the Gibbs free energy of light transition metal oxides.
We develop new transfer learning algorithms to accelerate prediction of material properties from ab initio simulations based on density functional theory (DFT). Transfer learning has been successfully utilized for data-efficient modeling in applications other than materials science, and it allows transferable representations learned from large datasets to be repurposed for learning new tasks even with small datasets. In the context of materials science, this opens the possibility to develop generalizable neural network models that can be repurposed on other materials, without the need of generating a large (computationally expensive) training set of materials properties. The proposed transfer learning algorithms are demonstrated on predicting the Gibbs free energy of light transition metal oxides.