Data and Code for Semantics-Audited and Physics-Constrained Multiview Learning for Cross-Database Prediction of Piezoelectric Response
收藏官方服务:
资源简介:
Reproducibility archive supporting the manuscript “Semantics-Audited and Physics-Constrained Multiview Learning for Cross-Database Prediction of Piezoelectric Response.” It contains derived Materials Project and JARVIS-DFT identifiers and tensor targets, formula- and prototype-disjoint predictions, the 39-record external validation cohort, the label-blind 120-record stress test, bootstrap and uncertainty outputs, semantic and leakage audits, candidate decisions, original analysis scripts, and editable manuscript figures. Original database download archives, API credentials, cached models, and third-party source-code copies are excluded. See README.md and MANIFEST_SHA256.csv for the directory map, provenance, and file checksums.
提供机构:
Zenodo创建时间:
2026-08-13



