A Benchmark of Rule-Based and Neural Coreference Resolution in Dutch Novels and News ...
We evaluate a rule-based (Lee et al., 2013) and neural (Lee et al., 2018) coreference system on Dutch datasets of two domains: literary novels and news/Wikipedia text. The results provide insight into the relative strengths of data-driven and knowledge-driven systems, as well as the influence of domain, document length, and annotation schemes. The neural system performs best on news/Wikipedia text, while the rule-based system performs best on literature. The neural system shows weaknesses with limited training data and long documents, while the rule-based system is affected by annotation diffe... Mehr ...
| Verfasser: | |
|---|---|
| Dokumenttyp: | Artikel |
| Erscheinungsdatum: | 2020 |
| Verlag/Hrsg.: |
arXiv
|
| Schlagwörter: | Computation and Language cs.CL / FOS: Computer and information sciences |
| Sprache: | unknown |
| Permalink: | https://search.fid-benelux.de/Record/base-29396190 |
| Datenquelle: | BASE; Originalkatalog |
| Powered By: | BASE |
| Link(s) : | https://dx.doi.org/10.48550/arxiv.2011.01615 |
We evaluate a rule-based (Lee et al., 2013) and neural (Lee et al., 2018) coreference system on Dutch datasets of two domains: literary novels and news/Wikipedia text. The results provide insight into the relative strengths of data-driven and knowledge-driven systems, as well as the influence of domain, document length, and annotation schemes. The neural system performs best on news/Wikipedia text, while the rule-based system performs best on literature. The neural system shows weaknesses with limited training data and long documents, while the rule-based system is affected by annotation differences. The code and models used in this paper are available at https://github.com/andreasvc/crac2020 ... : Accepted for CRAC 2020 @ COLING ...