Источник
AI4X
Дата публикации
08.07.2025
Авторы
Артем Дембицкий Роман Еремин Иннокентий Хумонен Семен Буденный Станислав Федотов Дмитрий Аксёнов
Поделиться

Datasets for Benchmarking Machine Learning Models for Accelerated Search of Fast Ionic Conductors

Аннотация

Fast ionic conductors are essential for advancing electrochemical devices, such as batteries, gas sensors, and ceramic membranes [1]. Traditional searches for these materials rely on costly highthroughput (HT) density functional theory (DFT) calculations across diverse chemical and structural spaces. To accelerate technological advancement in this area, HT schemes must be optimized by replacing DFTwith accurate and fast surrogate models for rapid materials screening. Here, we benchmark existing machine learning (ML) models for HT search of Li-ion conductors.




Присоединяйтесь к AIRI в соцсетях