Matscibert

🚀 MatSciBERT

A Materials Domain Language Model for Text Mining and Information Extraction. This model is trained on material science research papers to facilitate text mining and information extraction in the materials domain.

🚀 Quick Start

This is the pretrained model presented in MatSciBERT: A materials domain language model for text mining and information extraction, which is a BERT model trained on material science research papers.

✨ Features

• Metrics: The model is evaluated using metrics such as accuracy, f1, precision, and recall.
• Library: It is based on the transformers library.

📦 Installation

No specific installation steps are provided in the original document, so this section is skipped.

💻 Usage Examples

No code examples are provided in the original document, so this section is skipped.

📚 Documentation

The training corpus comprises papers related to the broad category of materials: alloys, glasses, metallic glasses, cement and concrete. We have utilised the abstracts and full text of papers(when available). All the research papers have been downloaded from ScienceDirect using the Elsevier API. The detailed methodology is given in the paper.

The codes for pretraining and finetuning on downstream tasks are shared on GitHub.

🔧 Technical Details

No in - depth technical details (more than 50 words of specific technical description) are provided in the original document, so this section is skipped.

📄 License

The model is released under the MIT license.

If you find this useful in your research, please consider citing:

@article{gupta_matscibert_2022,
  title   = "{MatSciBERT}: A Materials Domain Language Model for Text Mining and Information Extraction",
  author  = "Gupta, Tanishq and 
            Zaki, Mohd and 
            Krishnan, N. M. Anoop and 
            Mausam",
  year    = "2022",
  month   = may,
  journal = "npj Computational Materials",
  volume  = "8",
  number  = "1",
  pages   = "102",
  issn    = "2057-3960",
  url     = "https://www.nature.com/articles/s41524-022-00784-w",
  doi     = "10.1038/s41524-022-00784-w"
}