Sense Vocabulary Compression through the Semantic Knowledge of WordNet for Neural Word Sense Disambiguation

Loïc Vial, Benjamin Lecouteux, Didier Schwab


Abstract
In this article, we tackle the issue of the limited quantity of manually sense annotated corpora for the task of word sense disambiguation, by exploiting the semantic relationships between senses such as synonymy, hypernymy and hyponymy, in order to compress the sense vocabulary of Princeton WordNet, and thus reduce the number of different sense tags that must be observed to disambiguate all words of the lexical database. We propose two different methods that greatly reduce the size of neural WSD models, with the benefit of improving their coverage without additional training data, and without impacting their precision. In addition to our methods, we present a WSD system which relies on pre-trained BERT word vectors in order to achieve results that significantly outperforms the state of the art on all WSD evaluation tasks.
Anthology ID:
2019.gwc-1.14
Volume:
Proceedings of the 10th Global Wordnet Conference
Month:
July
Year:
2019
Address:
Wroclaw, Poland
Editors:
Piek Vossen, Christiane Fellbaum
Venue:
GWC
SIG:
SIGLEX
Publisher:
Global Wordnet Association
Note:
Pages:
108–117
Language:
URL:
https://aclanthology.org/2019.gwc-1.14
DOI:
Bibkey:
Cite (ACL):
Loïc Vial, Benjamin Lecouteux, and Didier Schwab. 2019. Sense Vocabulary Compression through the Semantic Knowledge of WordNet for Neural Word Sense Disambiguation. In Proceedings of the 10th Global Wordnet Conference, pages 108–117, Wroclaw, Poland. Global Wordnet Association.
Cite (Informal):
Sense Vocabulary Compression through the Semantic Knowledge of WordNet for Neural Word Sense Disambiguation (Vial et al., GWC 2019)
Copy Citation:
PDF:
https://aclanthology.org/2019.gwc-1.14.pdf
Code
 getalp/disambiguate +  additional community code
Data
Word Sense Disambiguation: a Unified Evaluation Framework and Empirical Comparison