Bayesian Hierarchical Words Representation Learning

Oren Barkan; Idan Rejwan; Avi Caciularu; Noam Koenigstein

doi:10.18653/v1/2020.acl-main.356

Bayesian Hierarchical Words Representation Learning

Oren Barkan, Idan Rejwan, Avi Caciularu, Noam Koenigstein

Department of Industrial Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

This paper presents the Bayesian Hierarchical Words Representation (BHWR) learning algorithm. BHWR facilitates Variational Bayes word representation learning combined with semantic taxonomy modeling via hierarchical priors. By propagating relevant information between related words, BHWR utilizes the taxonomy to improve the quality of such representations. Evaluation of several linguistic datasets demonstrates the advantages of BHWR over suitable alternatives that facilitate Bayesian modeling with or without semantic priors. Finally, we further show that BHWR produces better representations for rare words.

Original language	English
Title of host publication	58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020)
Editors	Dan Jurafsky, Joyce Chai, Natalie Schluter, Joel Tetreault
Place of Publication	209 N EIGHTH STREET, STROUDSBURG, PA 18360 USA
Publisher	ASSOC COMPUTATIONAL LINGUISTICS-ACL
Pages	3871-3877
Number of pages	7
ISBN (Print)	978-1-952148-25-5
DOIs	https://doi.org/10.18653/v1/2020.acl-main.356
State	Published - 2020

Access to Document

10.18653/v1/2020.acl-main.356

https://aclanthology.org/2020.acl-main.356.pdf

Cite this

@inproceedings{bd390eaa22134530a9ddf21d6da43698,

title = "Bayesian Hierarchical Words Representation Learning",

abstract = "This paper presents the Bayesian Hierarchical Words Representation (BHWR) learning algorithm. BHWR facilitates Variational Bayes word representation learning combined with semantic taxonomy modeling via hierarchical priors. By propagating relevant information between related words, BHWR utilizes the taxonomy to improve the quality of such representations. Evaluation of several linguistic datasets demonstrates the advantages of BHWR over suitable alternatives that facilitate Bayesian modeling with or without semantic priors. Finally, we further show that BHWR produces better representations for rare words.",

author = "Oren Barkan and Idan Rejwan and Avi Caciularu and Noam Koenigstein",

note = "58th Annual Meeting of the Association-for-Computational-Linguistics (ACL), ELECTR NETWORK, JUL 05-10, 2020",

year = "2020",

doi = "10.18653/v1/2020.acl-main.356",

language = "אנגלית",

isbn = "978-1-952148-25-5",

pages = "3871--3877",

editor = "Dan Jurafsky and Joyce Chai and Natalie Schluter and Joel Tetreault",

booktitle = "58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020)",

publisher = "ASSOC COMPUTATIONAL LINGUISTICS-ACL",

}

Barkan, O, Rejwan, I, Caciularu, A & Koenigstein, N 2020, Bayesian Hierarchical Words Representation Learning. in D Jurafsky, J Chai, N Schluter & J Tetreault (eds), 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020). ASSOC COMPUTATIONAL LINGUISTICS-ACL, 209 N EIGHTH STREET, STROUDSBURG, PA 18360 USA, pp. 3871-3877. https://doi.org/10.18653/v1/2020.acl-main.356

Bayesian Hierarchical Words Representation Learning. / Barkan, Oren; Rejwan, Idan; Caciularu, Avi et al.

58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020). ed. / Dan Jurafsky; Joyce Chai; Natalie Schluter; Joel Tetreault. 209 N EIGHTH STREET, STROUDSBURG, PA 18360 USA : ASSOC COMPUTATIONAL LINGUISTICS-ACL, 2020. p. 3871-3877.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Bayesian Hierarchical Words Representation Learning

AU - Barkan, Oren

AU - Rejwan, Idan

AU - Caciularu, Avi

AU - Koenigstein, Noam

N1 - 58th Annual Meeting of the Association-for-Computational-Linguistics (ACL), ELECTR NETWORK, JUL 05-10, 2020

PY - 2020

Y1 - 2020

N2 - This paper presents the Bayesian Hierarchical Words Representation (BHWR) learning algorithm. BHWR facilitates Variational Bayes word representation learning combined with semantic taxonomy modeling via hierarchical priors. By propagating relevant information between related words, BHWR utilizes the taxonomy to improve the quality of such representations. Evaluation of several linguistic datasets demonstrates the advantages of BHWR over suitable alternatives that facilitate Bayesian modeling with or without semantic priors. Finally, we further show that BHWR produces better representations for rare words.

AB - This paper presents the Bayesian Hierarchical Words Representation (BHWR) learning algorithm. BHWR facilitates Variational Bayes word representation learning combined with semantic taxonomy modeling via hierarchical priors. By propagating relevant information between related words, BHWR utilizes the taxonomy to improve the quality of such representations. Evaluation of several linguistic datasets demonstrates the advantages of BHWR over suitable alternatives that facilitate Bayesian modeling with or without semantic priors. Finally, we further show that BHWR produces better representations for rare words.

U2 - 10.18653/v1/2020.acl-main.356

DO - 10.18653/v1/2020.acl-main.356

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

SN - 978-1-952148-25-5

SP - 3871

EP - 3877

BT - 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020)

A2 - Jurafsky, Dan

A2 - Chai, Joyce

A2 - Schluter, Natalie

A2 - Tetreault, Joel

PB - ASSOC COMPUTATIONAL LINGUISTICS-ACL

CY - 209 N EIGHTH STREET, STROUDSBURG, PA 18360 USA

ER -

Barkan O, Rejwan I, Caciularu A, Koenigstein N. Bayesian Hierarchical Words Representation Learning. In Jurafsky D, Chai J, Schluter N, Tetreault J, editors, 58TH ANNUAL MEETING OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS (ACL 2020). 209 N EIGHTH STREET, STROUDSBURG, PA 18360 USA: ASSOC COMPUTATIONAL LINGUISTICS-ACL. 2020. p. 3871-3877 doi: 10.18653/v1/2020.acl-main.356

Bayesian Hierarchical Words Representation Learning

Abstract

Access to Document

Fingerprint

Cite this