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