Learning the Geometry of Materials Space with Trained Neural Tangent Kernels

Abstract

Neural networks trained to predict materials properties learn representations whose physical content can be difficult to extract. We use the neural tangent kernel (NTK), evaluated at trained parameters, to recover the geometry learned by a network that predicts superconducting transition temperatures from elemental composition alone. Kernel ridge regression with the trained NTK closely reproduces the network's predictive performance and substantially outperforms regression with the initial kernel. Spectral analysis and multidimensional scaling reveal a geometry dominated by a small number of directions. Clustering and local neighborhood analysis organize compounds into recognizable superconducting families and connect materials with related chemical and correlated-electron properties despite substantial differences in $T_c$. These results establish trained tangent kernels as practical probes of learned materials representations, connecting property prediction with the exploration of materials space.

The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.

More in Materials Science

• arXiv preprint · 21 October 2026

  High-fidelity and robust controlled-Z gates for superconducting transmon qubits. I. Intrinsic errors and decoherence

  Chia-Hsien Huang and Hsi-Sheng Goan
• arXiv preprint · 19 October 2026

  Sidewall Metrology Protocol for Superconducting Qubit Devices

  Jørgen Garnæs, Henrik S. Guttesen, Astrid T. Rømer and 9 others
• arXiv preprint · 21 October 2026

  Extreme competition between stripes and superconductivity in the hole-doped Hubbard model

  Ao Chen, Johnnie Gray, Si-Jing Du and 1 other
• arXiv preprint · 16 October 2026

  Experimental Demonstration of a Balanced Travelling Wave Parametric Amplifier

  Nikita Klimovich, Peter K. Day, Henry G. Leduc and 1 other
• arXiv preprint · 22 October 2026

  Cluster property and entanglement in the superconducting state

  Katsuhiko Higuchi and Masahiko Higuchi

All new papers in Materials Science