Beyond Molecular Structure: Critically Assessing Machine Learning for Designing Organic Photovoltaic Materials and Devices
Bibliographic record
Abstract
Our study explores the current state of machine learning (ML) as applied to predicting and designing organic photovoltaic (OPV) devices. We outline key considerations for selecting the method of encoding a molecular structure and selecting the algorithm while also emphasizing important aspects of training and rigorously evaluating ML models. This work presents the first dataset of OPV device fabrication data mined from the literature. The top models achieve state-of-the-art predictive performance. In particular, we identify an algorithm that is used less frequently, but may be particularly well suited to similar datasets. However, predictive performance remains modest (R2 ≅ 0.6) overall. An in-depth analysis of the dataset attributes this limitation to challenges relating to the size of the dataset, as well as data quality and sparsity. These aspects are directly tied to difficulties imposed by current reporting and publication practices. Advocating for standardized reporting of OPV device fabrication data reporting in publications emerges as crucial to streamline literature mining and foster ML adoption. This comprehensive investigation emphasizes the critical role of both data quantity and quality, and highlights the need for collective efforts to unlock ML's potential to drive advancements in OPV.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.027 | 0.098 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.003 | 0.004 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.006 | 0.009 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.003 | 0.004 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".