krict-data/simd

A Public Database of Thermoelectric Materials and System-Identified Material Representation for Data-Driven Discovery

Thermoelectric materials have received much attention for energy harvesting devices and power generators. However, discovering novel high-performance thermoelectric materials is a challenging task due to the diversity and the structural complexities of the thermoelectric materials containing alloys and dopants. For efficient data-driven discovery of novel thermoelectric materials, we constructed a public dataset that contains experimentally synthesized thermoelectric materials and their experimental thermoelectric properties. In our dataset, we achieved $R^2$-scores greater than 0.9 in the regression problems for predicting experimentally measured thermoelectric properties of the materials from their chemical compositions. Furthermore, we devised a material descriptor for the chemical compositions of the materials to improve extrapolation capabilities of machine learning methods. Based on transfer learning with the proposed material descriptor, we greatly improved $R^2$-score from 0.13 to 0.71 in predicting experimental ZTs of the materials from completely unseen material groups.

Reference: https://doi.org/10.1038/s41524-022-00897-2

Run

This repository provides an implementation of transfer learning based on System-Identified Material Representation (SIMD). By executing exec.py, you can train and evaluate the XGBoost regressor with SIMD to predict ZTs of thermoelectric materials from unexplored material groups.

Datasets

To reproduce the extrapolation results of SIMD, we should prepare the following two datasets of thermoelectric materials.

• Starry dataset: It is a large materials dataset containing thermoelectric materials. Since it was collected by text mining, data pre-processing should be conducted to remove invalid data (reference: https://www.starrydata2.org).
• ESTM dataset: It is a refined thermoelectric materials dataset for machine learning. ESTM dataset contains 5,205 experimental observations of thermoelectric materials and their properties (reference: https://doi.org/10.1038/s41524-022-00897-2).

Notes

• This repository contains only a subset of the source Starry dataset due to the dataset license. Please visit Starrydata to download the full data of the source Starry dataset.
• The full data of the ESTM dataset is provided in the dataset folder of this repository.
• The results folder provides the extrapolation results on the full data of the Starry and ESTM dataset. You can check the extrapolation results reported in the paper.

simd's Issues

No valid DOI for dataset

The current DOI on the README is given as:

https://doi.org/10.xxxx/xxxxxxxxx

Is there a valid DOI for this dataset? I don't see anything on the corresponding paper: https://doi.org/10.1038/s41524-022-00897-2