Exploring diamond-like lattice thermal conductivity crystals via feature-based transfer learning

TL;DR

This paper demonstrates how transfer learning using big data and feature-based neural networks can accurately predict high thermal conductivity in crystals, enabling efficient screening of over 60,000 compounds for diamond-like materials.

Contribution

It introduces a transfer learning approach that combines big and small data with feature descriptors to predict thermal conductivity and identify novel high-conductivity crystals.

Findings

Successful transfer learning enables extrapolative predictions.

Descriptors for lattice anharmonicity are revealed.

Over 60,000 compounds screened for potential diamond-like materials.

Abstract

Ultrahigh lattice thermal conductivity materials hold great importance since they play a critical role in the thermal management of electronic and optical devices. Models using machine learning can search for materials with outstanding higher-order properties like thermal conductivity. However, the lack of sufficient data to train a model is a serious hurdle. Herein we show that big data can complement small data for accurate predictions when lower-order feature properties available in big data are selected properly and applied to transfer learning. The connection between the crystal information and thermal conductivity is directly built with a neural network by transferring descriptors acquired through a pre-trained model for the feature property. Successful transfer learning shows the ability of extrapolative prediction and reveals descriptors for lattice anharmonicity. Transfer learning is employed to screen over 60000 compounds to identify novel crystals that can serve as alternatives to diamond.