PROTEST-ER: Retraining BERT for Protest Event Extraction
| Autor: | Caselli, Tommaso, Mutlu, O., Basile, Angelo, Hürriyetoğlu, A., Hürriyetoğlu, Ali |
|---|---|
| Přispěvatelé: | Hürriyetoğlu, Ali, Mutlu, Osman, Caselli, Tommaso, Basile, Angelo, College of Social Sciences and Humanities, College of Engineering, Department of Sociology, Department of Computer Engineering, Computational Linguistics (CL) |
| Rok vydání: | 2021 |
| Předmět: |
03 medical and health sciences
Domain adaptation Software portability 0302 clinical medicine Computer science Event (computing) Speech recognition Retraining Artificial intelligence Computer science interdisciplinary applications Linguistics 030212 general & internal medicine 030204 cardiovascular system & hematology Adaptation strategies Different domains Domain specific Events extractions Extraction modeling Performance Pre-training Text genre Domain (software engineering) |
| Zdroj: | Proceedings of the 4th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2021) Proceedings of the 4th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2021), 12-19 STARTPAGE=12;ENDPAGE=19;TITLE=Proceedings of the 4th Workshop on Challenges and Applications of Automated Extraction of Socio-political Events from Text (CASE 2021) |
| DOI: | 10.18653/v1/2021.case-1.4 |
| Popis: | We analyze the effect of further pre-training BERT with different domain specific data as an unsupervised domain adaptation strategy for event extraction. Portability of event extraction models is particularly challenging, with large performance drops affecting data on the same text genres (e.g., news). We present PROTEST-ER, a retrained BERT model for protest event extraction. PROTEST-ER outperforms a corresponding generic BERT on out-of-domain data of 8.1 points. Our best performing models reach 51.91-46.39 F1 across both domains. European Union (EU); Horizon 2020; European Research Council (ERC) |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|