Definition Frames: Using Definitions for Hybrid Concept Representations

Evangelia Spiliopoulou, Artidoro Pagnoni, Eduard Hovy


Abstract
Advances in word representations have shown tremendous improvements in downstream NLP tasks, but lack semantic interpretability. In this paper, we introduce Definition Frames (DF), a matrix distributed representation extracted from definitions, where each dimension is semantically interpretable. DF dimensions correspond to the Qualia structure relations: a set of relations that uniquely define a term. Our results show that DFs have competitive performance with other distributional semantic approaches on word similarity tasks.
Anthology ID:
2020.coling-main.273
Volume:
Proceedings of the 28th International Conference on Computational Linguistics
Month:
December
Year:
2020
Address:
Barcelona, Spain (Online)
Venue:
COLING
SIG:
Publisher:
International Committee on Computational Linguistics
Note:
Pages:
3060–3068
Language:
URL:
https://aclanthology.org/2020.coling-main.273
DOI:
10.18653/v1/2020.coling-main.273
Bibkey:
Cite (ACL):
Evangelia Spiliopoulou, Artidoro Pagnoni, and Eduard Hovy. 2020. Definition Frames: Using Definitions for Hybrid Concept Representations. In Proceedings of the 28th International Conference on Computational Linguistics, pages 3060–3068, Barcelona, Spain (Online). International Committee on Computational Linguistics.
Cite (Informal):
Definition Frames: Using Definitions for Hybrid Concept Representations (Spiliopoulou et al., COLING 2020)
Copy Citation:
PDF:
https://preview.aclanthology.org/ingestion-script-update/2020.coling-main.273.pdf
Code
 spilioeve/Definition-Frames