Model documentation & parameters
Parameters
Model
Whether to use the model trained 1) on procedures for heterogeneous single-atom catalyst synthesis, or 2) on organic chemistry procedures.
Synthesis text
Synthesis procedure (in English prose) to extract actions from.
Model card -- Text mining synthesis protocols of heterogeneous single-atom catalysts
Model Details: Sequence-to-sequence transformer model
Developers: Manu Suvarna, Alain C. Vaucher, Sharon Mitchell, Teodoro Laino, and Javier Pérez-Ramírez.
Distributors: Same as the developers.
Model date: April 2023.
Algorithm version: Details in the source code and in the paper.
Model type: A Transformer-based sequence-to-sequence language model that extracts synthesis actions from procedure text. The model relies on the OpenNMT library.
Information about training algorithms, parameters, fairness constraints or other applied approaches, and features: Details in the source code and in the paper.
Paper or other resource for more information: Currently under review.
License: MIT
Where to send questions or comments about the model: Contact one of the developers.
Intended Use. Use cases that were envisioned during development: Chemical research, in particular in the field of heterogeneous single-atom catalysts.
Primary intended uses/users: Researchers and computational chemists using the model for model comparison or research exploration purposes.
Out-of-scope use cases: Production-level inference.
Factors: Not applicable.
Metrics: Details in the source code and in the paper.
Datasets: Details in the source code and in the paper.
Ethical Considerations: No specific considerations as no private/personal data is involved. Please consult with the authors in case of questions.
Caveats and Recommendations: Please consult with original authors in case of questions.
Model card prototype inspired by Mitchell et al. (2019).
Citation
@article{suvarna2023textmining,
title={Text mining and standardization of single-atom catalyst protocols to foster digital synthesis},
author={Manu Suvarna, Alain C. Vaucher, Sharon Mitchell, Teodoro Laino, and Javier Pérez-Ramírez},
journal={under review},
}