feradauto/scibert_nlp4sg

SciBERT NLP4SG

SciBERT NLP4SG is a SciBERT model fine-tuned to detect NLP4SG papers based on their title and abstract.

We present the details in the paper:

The training corpus is a combination of the NLP4SGPapers training set which is manually annotated, and some papers identified by keywords.

For more details about the training data and the model, visit the original repo here.

Please cite the following paper:

@misc{gonzalez2023good,
      title={Beyond Good Intentions: Reporting the Research Landscape of NLP for Social Good}, 
      author={Fernando Gonzalez and Zhijing Jin and Jad Beydoun and Bernhard Schölkopf and Tom Hope and Mrinmaya Sachan and Rada Mihalcea},
      year={2023},
      eprint={2305.05471},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}