Testing APSyn against vector cosine on similarity estimation

Enrico Santus; Emmanuele Chersoni; Alessandro Lenci; Chu-ren Huang; Philippe Blache

Testing APSyn against vector cosine on similarity estimation

Enrico Santus, Emmanuele Chersoni, Alessandro Lenci, Chu-ren Huang, Philippe Blache

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

10 Citations (Scopus)

Abstract

In Distributional Semantic Models (DSMs), Vector Cosine is widely used to estimate similarity between word vectors, although this measure was noticed to suffer from several shortcomings. The recent lit ENGLerature has proposed other methods which attempt to mitigate such biases. In this paper, we intend to investigate APSyn, a measure that computes the extent of the intersection between the most associated contexts of two target words, weighting it by context relevance. We evaluated this metric in a similarity estimation task on several popular test sets, and our results show that APSyn is in fact highly competitive, even with respect to the results reported in the lit ENGLerature for word embeddings. On top of it, APSyn addresses some of the weaknesses of Vector Cosine, performing well also on genuine similarity estimation.

Original language	English
Title of host publication	Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016
Publisher	Institute for the Study of Language and Information
Pages	229-238
Number of pages	10
ISBN (Electronic)	9788968174285
Publication status	Published - 1 Jan 2016
Event	30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016 - Kyung Hee University, Seoul, Korea, Republic of Duration: 28 Oct 2016 → 30 Oct 2016

Conference

Conference	30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016
Country/Territory	Korea, Republic of
City	Seoul
Period	28/10/16 → 30/10/16

ASJC Scopus subject areas

Language and Linguistics
Computer Science (miscellaneous)
Information Systems

More information

http://hdl.handle.net/10397/71889

Cite this

@inproceedings{4b248863ddae4e69bad4abeb18e3fcf9,

title = "Testing APSyn against vector cosine on similarity estimation",

abstract = "In Distributional Semantic Models (DSMs), Vector Cosine is widely used to estimate similarity between word vectors, although this measure was noticed to suffer from several shortcomings. The recent lit ENGLerature has proposed other methods which attempt to mitigate such biases. In this paper, we intend to investigate APSyn, a measure that computes the extent of the intersection between the most associated contexts of two target words, weighting it by context relevance. We evaluated this metric in a similarity estimation task on several popular test sets, and our results show that APSyn is in fact highly competitive, even with respect to the results reported in the lit ENGLerature for word embeddings. On top of it, APSyn addresses some of the weaknesses of Vector Cosine, performing well also on genuine similarity estimation.",

author = "Enrico Santus and Emmanuele Chersoni and Alessandro Lenci and Chu-ren Huang and Philippe Blache",

year = "2016",

month = jan,

day = "1",

language = "English",

pages = "229--238",

booktitle = "Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016",

publisher = "Institute for the Study of Language and Information",

note = "30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016 ; Conference date: 28-10-2016 Through 30-10-2016",

}

Santus, E, Chersoni, E, Lenci, A, Huang, C & Blache, P 2016, Testing APSyn against vector cosine on similarity estimation. in Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016. Institute for the Study of Language and Information, pp. 229-238, 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016, Seoul, Korea, Republic of, 28/10/16. <http://hdl.handle.net/10397/71889>

Testing APSyn against vector cosine on similarity estimation. / Santus, Enrico; Chersoni, Emmanuele; Lenci, Alessandro et al.
Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016. Institute for the Study of Language and Information, 2016. p. 229-238.

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Testing APSyn against vector cosine on similarity estimation

AU - Santus, Enrico

AU - Chersoni, Emmanuele

AU - Lenci, Alessandro

AU - Huang, Chu-ren

AU - Blache, Philippe

PY - 2016/1/1

Y1 - 2016/1/1

N2 - In Distributional Semantic Models (DSMs), Vector Cosine is widely used to estimate similarity between word vectors, although this measure was noticed to suffer from several shortcomings. The recent lit ENGLerature has proposed other methods which attempt to mitigate such biases. In this paper, we intend to investigate APSyn, a measure that computes the extent of the intersection between the most associated contexts of two target words, weighting it by context relevance. We evaluated this metric in a similarity estimation task on several popular test sets, and our results show that APSyn is in fact highly competitive, even with respect to the results reported in the lit ENGLerature for word embeddings. On top of it, APSyn addresses some of the weaknesses of Vector Cosine, performing well also on genuine similarity estimation.

AB - In Distributional Semantic Models (DSMs), Vector Cosine is widely used to estimate similarity between word vectors, although this measure was noticed to suffer from several shortcomings. The recent lit ENGLerature has proposed other methods which attempt to mitigate such biases. In this paper, we intend to investigate APSyn, a measure that computes the extent of the intersection between the most associated contexts of two target words, weighting it by context relevance. We evaluated this metric in a similarity estimation task on several popular test sets, and our results show that APSyn is in fact highly competitive, even with respect to the results reported in the lit ENGLerature for word embeddings. On top of it, APSyn addresses some of the weaknesses of Vector Cosine, performing well also on genuine similarity estimation.

UR - http://www.scopus.com/inward/record.url?scp=85015921275&partnerID=8YFLogxK

M3 - Conference article published in proceeding or book

SP - 229

EP - 238

BT - Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016

PB - Institute for the Study of Language and Information

T2 - 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016

Y2 - 28 October 2016 through 30 October 2016

ER -

Testing APSyn against vector cosine on similarity estimation

Abstract

Conference

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this