Abstract
Universal Dependencies incur a high cost in computation for unbiased system development. We propose a 100% empirically chosen small subset of UD languages for efficient parsing system development. The technique used is based on measurements of model capacity globally. We show that the diversity of the resulting representative language set is superior to the requirements-based procedure.
| Original language | English |
|---|---|
| Conference proceedings | NEALT (Northern European Association of Language Technology) Proceedings Series |
| Volume | 31 |
| Pages (from-to) | 117-122 |
| ISSN | 1736-6305 |
| Publication status | Published - 2017 |
| Event | NoDaLiDa 2017 Workshop on Universal Dependencies - Gothenburg, Sweden Duration: 22 May 2017 → … |
Workshop
| Workshop | NoDaLiDa 2017 Workshop on Universal Dependencies |
|---|---|
| Country/Territory | Sweden |
| City | Gothenburg |
| Period | 22/05/2017 → … |
Keywords
- Universal Dependencies
- Parsers
- Language subset selection
- Model capacity
- Empirical methods
Fingerprint
Dive into the research topics of 'Empirically sampling Universal Dependencies'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver