Machine learning reveals common features of unconventional superconductors with high transition temperatures
Abstract
Superconductors with high critical temperatures that emerges beyond the phonon-mediated regime are usually considered unconventional in nature, yet unlike conventional superconductors, no broadly applicable predictive theory currently guides their discovery. Here, we use interpretable machine learning to uncover a common materials-space signature of high-Tc unconventional superconductors and develop a data-driven strategy for materials discovery. We construct a unified feature representation for each material by integrating compositional statistics, structural information, and latent representations from trained property-prediction models, followed by structure-aware filtering of an experimentally established superconducting dataset. Without using transition-temperature information, unsupervised analysis shows that cuprate and iron-based superconductors occupy a common region of materials space, characterized primarily by large electronegativity deviation and intermediate mean valence-electron number. A supervised Tc model independently identifies the same descriptors as dominant features, providing complementary evidence for their relevance. Using this empirical materials-space prior together with the Tc model, we prioritize candidate materials, recover recently discovered nickelate superconductors, and identify chemically distinct candidates for future investigation.
Related
- Gaps in unconventional superconductors
- Resolving unconventional gap structure in kagome superconductors with hybrid microwave circuits
- Pre-pairing and polarization of doped charge carriers in unconventional superconductors
- Chiral superconductors and competing states across a Lifshitz transition in rhombohedral pentalayer graphene
- Cryothermal Measurements of Variable-Emittance Coatings with Lower Phase Transition Temperatures for Space Thermal Control
BibTeX
@misc{xu2026machine,
title = {{Machine learning reveals common features of unconventional superconductors with high transition temperatures}},
author = {Haosheng Xu and Dongheng Qian and Yijun Yu and Jing Wang},
howpublished = {arXiv},
year = {2026},
month = aug,
doi = {10.48550/arxiv.2608.30588},
eprint = {2608.30588},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2608.30588},
}
RIS
TY - UNPB
TI - Machine learning reveals common features of unconventional superconductors with high transition temperatures
AU - Haosheng Xu
AU - Dongheng Qian
AU - Yijun Yu
AU - Jing Wang
PY - 2026
DA - 2026/08/31
PB - arXiv
DO - 10.48550/arxiv.2608.30588
UR - https://arxiv.org/abs/2608.30588
AB - Superconductors with high critical temperatures that emerges beyond the phonon-mediated regime are usually considered unconventional in nature, yet unlike conventional superconductors, no broadly applicable predictive theory currently guides their discovery. Here, we use interpretable machine learning to uncover a common materials-space signature of high-T_c unconventional superconductors and develop a data-driven strategy for materials discovery. We construct a unified feature representation for each material by integrating compositional statistics, structural information, and latent representations from trained property-prediction models, followed by structure-aware filtering of an experimentally established superconducting dataset. Without using transition-temperature information, unsupervised analysis shows that cuprate and iron-based superconductors occupy a common region of materials space, characterized primarily by large electronegativity deviation and intermediate mean valence-electron number. A supervised T_c model independently identifies the same descriptors as dominant features, providing complementary evidence for their relevance. Using this empirical materials-space prior together with the T_c model, we prioritize candidate materials, recover recently discovered nickelate superconductors, and identify chemically distinct candidates for future investigation.
ER -