The (too Many) Problems of Analogical Reasoning with Word Vectors

Anna Rogers, Aleksandr Drozd, Bofang Li


Abstract
This paper explores the possibilities of analogical reasoning with vector space models. Given two pairs of words with the same relation (e.g. man:woman :: king:queen), it was proposed that the offset between one pair of the corresponding word vectors can be used to identify the unknown member of the other pair (king - man + woman = queen). We argue against such “linguistic regularities” as a model for linguistic relations in vector space models and as a benchmark, and we show that the vector offset (as well as two other, better-performing methods) suffers from dependence on vector similarity.
Anthology ID:
S17-1017
Volume:
Proceedings of the 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017)
Month:
August
Year:
2017
Address:
Vancouver, Canada
Editors:
Nancy Ide, Aurélie Herbelot, Lluís Màrquez
Venue:
*SEM
SIGs:
SIGLEX | SIGSEM
Publisher:
Association for Computational Linguistics
Note:
Pages:
135–148
Language:
URL:
https://aclanthology.org/S17-1017
DOI:
10.18653/v1/S17-1017
Bibkey:
Cite (ACL):
Anna Rogers, Aleksandr Drozd, and Bofang Li. 2017. The (too Many) Problems of Analogical Reasoning with Word Vectors. In Proceedings of the 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017), pages 135–148, Vancouver, Canada. Association for Computational Linguistics.
Cite (Informal):
The (too Many) Problems of Analogical Reasoning with Word Vectors (Rogers et al., *SEM 2017)
Copy Citation:
PDF:
https://aclanthology.org/S17-1017.pdf