Information Type Classification with Contrastive Task-Specialized Sentence Encoders

  • User-generated information content has become an important information source in crisis situations. However, classification models suffer from noise and event-related biases which still poses a challenging task and requires sophisticated task-adaptation. To address these challenges, we propose the use of contrastive task-specialized sentence encoders for downstream classification. We apply the task-specialization on the CrisisLex, HumAID, and TrecIS information type classification tasks and show performance gains w.r.t. F1-score. Furthermore, we analyse the cross-corpus and cross-lingual capabilities for two German event relevancy classification datasets.

Export metadata

Additional Services

Search Google Scholar
Metadaten
Author:Philipp Seeberger, Tobias Bocklet, Korbinian Riedhammer
DOI:https://doi.org/10.48550/arXiv.2312.11020
ArXiv Id:http://arxiv.org/abs/2312.11020
Publisher:Association for Computational Linguistics
Document Type:conference proceeding (article)
Language:English
Date of first Publication:2023/12/18
Reviewed:Begutachtet/Reviewed
Release Date:2024/07/08
Pagenumber:180-186
Konferenzangabe:19th Conference on Natural Language Processing (KONVENS 2023)
institutes:Fakultät Informatik
Licence (German):Keine Lizenz - Deutsches Urheberrecht gilt
Verstanden ✔
Diese Webseite verwendet technisch erforderliche Session-Cookies. Durch die weitere Nutzung der Webseite stimmen Sie diesem zu. Unsere Datenschutzerklärung finden Sie hier.