Interpretable Word Embeddings via Informative Priors

Miriam Hurtado Bodell; Martin Arvidsson; Måns Magnusson

Interpretable Word Embeddings via Informative Priors

Miriam Hurtado Bodell, Martin Arvidsson, Måns Magnusson

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference contribution › Scientific › vertaisarvioitu

Abstrakti

Word embeddings have demonstrated strong performance on NLP tasks. However, lack of interpretability and the unsupervised nature of word embeddings have limited their use within computational social science and digital humanities. We propose the use of informative priors to create interpretable and domain-informed dimensions for probabilistic word embeddings. Experimental results show that sensible priors can capture latent semantic concepts better than or on-par with the current state of the art, while retaining the simplicity and generalizability of using priors.

Alkuperäiskieli	Englanti
Otsikko	The 2019 Conference on Empirical Methods in Natural Language Processing And the 9th International Joint Conference on Natural Language Processing
Alaotsikko	Proceedings of System Demonstrations
Sivut	6324-6330
Tila	Julkaistu - 2019
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisuussa
Tapahtuma	Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing - Hong Kong, Kiina Kesto: 3 marraskuuta 2019 → 7 marraskuuta 2019

Conference

Conference	Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing
Lyhennettä	MNLP/IJCNLP
Maa/Alue	Kiina
Kaupunki	Hong Kong
Ajanjakso	03/11/2019 → 07/11/2019

Pääsy asiakirjaan

https://arxiv.org/abs/1909.01459

Siteeraa tätä

@inproceedings{b47f20a06abb4a4cab2b1f7c35c14f9e,

title = "Interpretable Word Embeddings via Informative Priors",

abstract = "Word embeddings have demonstrated strong performance on NLP tasks. However, lack of interpretability and the unsupervised nature of word embeddings have limited their use within computational social science and digital humanities. We propose the use of informative priors to create interpretable and domain-informed dimensions for probabilistic word embeddings. Experimental results show that sensible priors can capture latent semantic concepts better than or on-par with the current state of the art, while retaining the simplicity and generalizability of using priors.",

author = "Bodell, {Miriam Hurtado} and Martin Arvidsson and M{\aa}ns Magnusson",

year = "2019",

language = "English",

isbn = "978-1-950737-92-5",

pages = "6324--6330",

booktitle = "The 2019 Conference on Empirical Methods in Natural Language Processing And the 9th International Joint Conference on Natural Language Processing",

note = "Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, MNLP/IJCNLP ; Conference date: 03-11-2019 Through 07-11-2019",

}

Bodell, MH, Arvidsson, M & Magnusson, M 2019, Interpretable Word Embeddings via Informative Priors. julkaisussa The 2019 Conference on Empirical Methods in Natural Language Processing And the 9th International Joint Conference on Natural Language Processing: Proceedings of System Demonstrations. Sivut 6324-6330, Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, Hong Kong, Kiina, 03/11/2019. <https://arxiv.org/abs/1909.01459>

Interpretable Word Embeddings via Informative Priors. / Bodell, Miriam Hurtado; Arvidsson, Martin; Magnusson, Måns.

The 2019 Conference on Empirical Methods in Natural Language Processing And the 9th International Joint Conference on Natural Language Processing: Proceedings of System Demonstrations. 2019. s. 6324-6330.

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference contribution › Scientific › vertaisarvioitu

TY - GEN

T1 - Interpretable Word Embeddings via Informative Priors

AU - Bodell, Miriam Hurtado

AU - Arvidsson, Martin

AU - Magnusson, Måns

PY - 2019

Y1 - 2019

N2 - Word embeddings have demonstrated strong performance on NLP tasks. However, lack of interpretability and the unsupervised nature of word embeddings have limited their use within computational social science and digital humanities. We propose the use of informative priors to create interpretable and domain-informed dimensions for probabilistic word embeddings. Experimental results show that sensible priors can capture latent semantic concepts better than or on-par with the current state of the art, while retaining the simplicity and generalizability of using priors.

AB - Word embeddings have demonstrated strong performance on NLP tasks. However, lack of interpretability and the unsupervised nature of word embeddings have limited their use within computational social science and digital humanities. We propose the use of informative priors to create interpretable and domain-informed dimensions for probabilistic word embeddings. Experimental results show that sensible priors can capture latent semantic concepts better than or on-par with the current state of the art, while retaining the simplicity and generalizability of using priors.

M3 - Conference contribution

SN - 978-1-950737-92-5

SP - 6324

EP - 6330

BT - The 2019 Conference on Empirical Methods in Natural Language Processing And the 9th International Joint Conference on Natural Language Processing

T2 - Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing

Y2 - 3 November 2019 through 7 November 2019

ER -

Interpretable Word Embeddings via Informative Priors

Abstrakti

Conference

Pääsy asiakirjaan

Sormenjälki

Siteeraa tätä