Provenance-grouped protocol for evaluating machine learning in perovskite photovoltaics

Key points

• Because nearly two-thirds of curated devices share their exact recipe with another, random splits primarily reward the mere recall of historical data rather than testing true physical generalisation.1
• To solve this, we mapped the institutional provenance of the Perovskite Database and established a rigorous hierarchy of grouped splits that progressively isolate duplicate devices, publications, and institutions.1
• Benchmarking thirteen models establishes new state-of-the-art performance across all splits and reveals a critical regime shift: while tree ensembles dominate random splits by recognising familiar recipes, their ranking collapses on genuinely unseen institutions and future discoveries, where composition-aware neural architectures excel.1
• Furthermore, feature importance demonstrates that the device architecture and fabrication methodology dictate final efficiency just as strongly as the absorber chemistry.1

Sources (1)

• [1]Provenance-grouped protocol for evaluating machine learning in perovskite photovoltaics
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 8, 10:17 PM

  Machine-learning models for perovskite solar cells are typically evaluated using random data splits, which we show systematically leak information through massive internal duplication.

  Because nearly two-thirds of curated devices share their exact recipe with another, random splits primarily reward the mere recall of historical data rather than testing true physical generalisation.

Extractive summary: sentences quoted from the sources.