Traversal-Free Word Vector Evaluation in Analogy Space

Xiaoyin Che, Nico Ring, Willi Raschkowski, Haojin Yang, Christoph Meinel


Abstract
In this paper, we propose an alternative evaluating metric for word analogy questions (A to B is as C to D) in word vector evaluation. Different from the traditional method which predicts the fourth word by the given three, we measure the similarity directly on the “relations” of two pairs of given words, just as shifting the relation vectors into a new analogy space. Cosine and Euclidean distances are then calculated as measurements. Observation and experiments shows the proposed analogy space evaluation could offer a more comprehensive evaluating result on word vectors with word analogy questions. Meanwhile, computational complexity are remarkably reduced by avoiding traversing the vocabulary.
Anthology ID:
W17-5302
Volume:
Proceedings of the 2nd Workshop on Evaluating Vector Space Representations for NLP
Month:
September
Year:
2017
Address:
Copenhagen, Denmark
Editors:
Samuel Bowman, Yoav Goldberg, Felix Hill, Angeliki Lazaridou, Omer Levy, Roi Reichart, Anders Søgaard
Venue:
RepEval
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
11–15
Language:
URL:
https://aclanthology.org/W17-5302
DOI:
10.18653/v1/W17-5302
Bibkey:
Cite (ACL):
Xiaoyin Che, Nico Ring, Willi Raschkowski, Haojin Yang, and Christoph Meinel. 2017. Traversal-Free Word Vector Evaluation in Analogy Space. In Proceedings of the 2nd Workshop on Evaluating Vector Space Representations for NLP, pages 11–15, Copenhagen, Denmark. Association for Computational Linguistics.
Cite (Informal):
Traversal-Free Word Vector Evaluation in Analogy Space (Che et al., RepEval 2017)
Copy Citation:
PDF:
https://aclanthology.org/W17-5302.pdf