BnPC: A Gold Standard Corpus for Paraphrase Detection in Bangla
Published in BUCC Workshop @ LREC-COLING 2024, 2024
| First Author | BUCC Workshop @ LREC-COLING 2024 | Citations: 4 |
- Automated collection of news headlines for paraphrase identification.
- Designed annotation schema, trained annotators, and benchmarked dataset with LMs and GPT models.
- Led a cross-dataset analysis proving the superiority of naturally occurring data over machine-translated data.
Recommended citation: Saha, S. et al. (2024). "BnPC: A Gold Standard Corpus for Paraphrase Detection in Bangla." BUCC Workshop @ LREC-COLING 2024.