Interpretable Word Embeddings via Informative Priors

Miriam Hurtado Bodell, Martin Arvidsson, Måns Magnusson

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaConference contributionScientificvertaisarvioitu

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.
AlkuperäiskieliEnglanti
OtsikkoThe 2019 Conference on Empirical Methods in Natural Language Processing And the 9th International Joint Conference on Natural Language Processing
AlaotsikkoProceedings of System Demonstrations
Sivut6324-6330
TilaJulkaistu - 2019
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisuussa
TapahtumaConference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing - Hong Kong, Kiina
Kesto: 3 marraskuuta 20197 marraskuuta 2019

Conference

ConferenceConference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing
LyhennettäMNLP/IJCNLP
MaaKiina
KaupunkiHong Kong
Ajanjakso03/11/201907/11/2019

Sormenjälki Sukella tutkimusaiheisiin 'Interpretable Word Embeddings via Informative Priors'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä