Abstract
Little evaluation has been performed on the many modeling options for span-based approaches. This paper investigates the performances of a wide range of span and context representation methods and their combinations with a focus on scientific named entity recognition (science NER). While some most common classical span encodings and their combination prove to be effective, few conclusions can be derived to context representations.
Original language | English |
---|---|
Title of host publication | Linking Theory and Practice of Digital Libraries |
Subtitle of host publication | 25th International Conference on Theory and Practice of Digital Libraries, TPDL 2021, Virtual Event, September 13–17, 2021, Proceedings |
Editors | Gerd Berget, Mark Michael Hall, Daniel Brenn, Sanna Kumpulainen |
Publisher | Springer |
Pages | 41-48 |
Number of pages | 8 |
ISBN (Electronic) | 978-3-030-86324-1 |
ISBN (Print) | 978-3-030-86323-4 |
DOIs | |
Publication status | Published - 7 Sept 2021 |
Event | 25th International Conference on Theory and Practice of Digital Libraries - Online, United Kingdom Duration: 13 Sept 2021 → 17 Sept 2021 Conference number: 25 http://www.tpdl.eu/tpdl2021/ |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 12866 LNCS |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 25th International Conference on Theory and Practice of Digital Libraries |
---|---|
Country/Territory | United Kingdom |
Period | 13/09/21 → 17/09/21 |
Internet address |
Keywords
- Context representation
- SciBERT
- Scientific named entity recognition
- Span representation
- Span-based model
ASJC Scopus subject areas
- Theoretical Computer Science
- General Computer Science