Interpretable Machine Learning for Materials Design

TL;DR

This paper compares different interpretable machine learning models for predicting materials properties, highlighting the trade-offs between interpretability and performance, and discusses future challenges and solutions in materials discovery.

Contribution

It evaluates four state-of-the-art ML techniques for materials property prediction and discusses strategies to maintain interpretability in complex models.

Findings

XGBoost, SISSO, Roost, and TPOT show varying trade-offs between accuracy and interpretability.

Challenges increase with larger datasets and more complex models.

Proposed solutions aim to enhance interpretability without sacrificing performance.

Abstract

Fueled by the widespread adoption of Machine Learning (ML) and the high-throughput screening of materials, the data-centric approach to materials design has asserted itself as a robust and powerful tool for the in-silico prediction of materials properties. When training models to predict material properties, researchers often face a difficult choice between a model's interpretability or its performance. We study this trade-off by leveraging four different state-of-the-art ML techniques: XGBoost, SISSO, Roost, and TPOT for the prediction of structural and electronic properties of perovskites and 2D materials. We then assess the future outlook of the continued integration of ML into materials discovery and identify key problems that will continue to challenge researchers as the size of the literature's datasets and complexity of models increases. Finally, we offer several possible solutions to these challenges with a focus on retaining interpretability and share our thoughts on magnifying the impact of ML on materials design.