Multilingual update! Check mREBEL, a multilingual version covering more relation types, languages and including entity types. REBEL : Relation Extraction By End to end Language generation This is the model card for the Findings of EMNLP 2021 paper REBEL: Relation Extraction By End to end Language generation. We present a new linearization approach and a reframing of Relation Extraction as a seq2seq task. The paper can be found here. If you use the code, please reference this work in your paper: @inproceedings{huguet cabot navigli 2021 rebel relation, title = "{REBEL}: Relation Extraction By End to end Language generation", author = "Huguet Cabot, Pere Llu{\'\i}s and Navigli, Roberto", booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021", month = nov, year = "2021", address = "Punta Cana, Dominican Republic", publisher = "Association for Computational Linguistics", url = "https://aclanthology.org/2021.findings emnlp.204", pages = "2370 2381", abstract = "Extracting relation triplets from raw text is a crucial task in Information Extraction, enabling multiple applications such as populating or validating knowledge bases, factchecking, and other downstre…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy