Crystal twins: self-supervised learning for crystalline material property prediction

Rishikesh Magar, Yuyang Wang, Amir Barati Farimani

npj Computational Materials 2022 article cited by 72

Source Open-access PDF More in Engineering & Materials

importance 6/10 ai assisted Machine Learning in Materials Science materials science stage: research

Problem

Accurate machine-learning prediction of crystalline material properties is limited by the scarcity of large labeled datasets, which are computationally expensive to generate.

Why it matters

Better property prediction reduces the cost of discovering new materials for technology and industry.

What the AI did

Introduced Crystal Twins, a self-supervised method using a twin Graph Neural Network that learns from unlabeled crystal data and improves fine-tuned downstream prediction.

Results

not stated in source

Impact

Abstract claims significantly improved downstream performance, but the magnitude is cut off and not stated.

Record

Venue	npj Computational Materials
Published	2022
Work type	article
Authors	Rishikesh Magar, Yuyang Wang, Amir Barati Farimani
Citations	72 (OpenAlex, at import time)
References	60
Model or method	Crystal Twins (twin GNN; Barlow Twins and SimSiam frameworks)
Organization	not stated in source
Primary source	https://doi.org/10.1038/s41524-022-00921-5
Open access	gold
Keywords	self-supervised learning, graph neural network, crystalline materials, property prediction, Barlow Twins

Importance rating

6 / 10

significant — thousands to millions affected

Property prediction underpins materials discovery for technology and industry; method improves results on 14 benchmarks, but abstract gives no scale of affected applications.

Basis: inferred. Judged from the source text, which gives no explicit scale. Stated scale: not stated in source. Method and limits.

More in Engineering & Materials

#	Paper	Year	Citations	Importance
1	Scaling deep learning for materials discovery Amil Merchant, Simon Batzner, Samuel S. Schoenholz, Muratahan Aykol, G...	2023	1,471	8
2	Application of Machine Learning in Material Synthesis and Property Prediction Guannan Huang, Yani Guo, Ye Chen, Zhengwei Nie	2023	165	7
3	Machine learning-guided discovery of ionic polymer electrolytes for lithium metal batteries Kai Li, Jifeng Wang, Yuanyuan Song, Ying Wang	2023	115	7
4	Machine Learning for Structural Health Monitoring of Aerospace Structures: A Review Gennaro Scarselli, Francesco Nicassio	2025	44	7
5	Power System Transient Stability Assessment Based on Machine Learning Algorithms and Grid Topology Mihail D. Senyuk, Murodbek Safaraliev, Firuz Kamalov, Hana Sulieman	2023	42	7