Work

Machine learning as a tool for predicting and engineering the next generation of piezoelectric materials

Year: 2026

Type: article

Abstract: Piezoelectric materials, which linearly interconvert mechanical and electrical energy, are ubiquitous in the modern world due to their many forms as sensors, actuators, and energy harvesters. Machine learning methods are emerging as a tool in the search for sustainable high-performance materials, offering the ability to rapidly analyze huge compositional spaces that are beyond traditional trial-and-error exploration. Recently, there has been a marked increase in the rate of publishing rationally designed piezoelectrics, especially methods that integrate machine learning. This review analyzes these recent developments and provides a meta-analysis in the field of machine learning predictions of piezoelectric properties. Collective issues such as limited data reporting and a lack of standardized evaluation metrics are highlighted as roadblocks to meaningful progress in high-accuracy predictions. Finally, some recommendations are made on best-practice methods to remedy these issues going forward—such as more open data sharing as well as standardized use of rigorous model validation schemes

Source: Cell Reports Physical Science

DOI: 10.1016/j.xcrp.2026.103522

Authors: Robert Guest, Sarah Guerin

Language: English

Topic: Machine Learning in Materials Science

Subfield: Materials Chemistry

Field: Materials Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $5,620