Discovery of interpretable Tc descriptors in conventional superconductors guided by symbolic regression
Fang Han Lim, Jinbo Pan, Shixuan Du
DOI 10.1103/4pxk-z8ww · Physical Review Materials
Active bibliographic source — not scientific approval
Bibliographic access preserves source history; it does not approve extracted materials or validate reported claims. Review warnings on each occurrence separately.
Abstract
Advancement of superconductivity is hindered by complexity of underlying mechanisms. Machine learning models have potential to accelerate materials discovery but face data limitations and interpretability challenges. This study identifies material genes governing conventional superconductivity through systematic two-stage analysis, where atomic descriptors are directly related to their phonon-mediated mechanism, unlike unconventional superconductors where strong electronic correlations that go beyond the Bardeen-Cooper-Schrieffer paradigm dominate in the formation of pairing glue. We first apply the random forest classifier and regressor to 16320 superconductors from NIMS database, screening over 114 MAGPIE and MEREDIG atomic descriptors to identify the 30 most important features. Combining these atomic descriptors with DFT-calculated bulk properties, we then perform SISSO (sure independence screening and sparsifying operator) symbolic regression on 60 carefully selected conventional superconductors. The resulting three-dimensional model [R2=0.778, AFD (average factor difference) =1.173] reveals optimal Tc requires an average of half filled or near half filled d electrons with moderate unfilled orbital heterogeneity. The discovered descriptors integrate d-valence electrons, unfilled orbitals, and electronegativity variations, providing actionable guidelines for materials design.
Similar papers
Accelerating the Search for Superconductors Using Machine Learning
similarity 0.89Suhas Adiga & Umesh V. Waghmare · 2025 · arXiv:2505.11964
Source status unknown — claims are unverified
Machine learning reveals common features of unconventional superconductors with high transition temperatures
similarity 0.89Haosheng Xu et al. · 2026 · arXiv:2608.30588
Source status unknown — claims are unverified
Machine-learning approach for discovery of conventional superconductors
similarity 0.88Huan Tran & Tuoc N. Vu · 2022 · arXiv:2211.03265
Source status unknown — claims are unverified
Functional form of the superconducting critical temperature from machine learning
similarity 0.88S. R. Xie et al.
Source status unknown — claims are unverified
Machine learning modeling of superconducting critical temperature
similarity 0.88Valentin Stanev et al. · 2017 · arXiv:1709.02727
Source status unknown — claims are unverified
AI-Driven Discovery of High-Temperature Superconductors via Materials Genome Initiative and High-Throughput Screening
similarity 0.88H. Gashmard et al. · 2025 · arXiv:2511.03865
Source status unknown — claims are unverified