In this paper, we propose a novel neural network model called RNN Encoder-Decoder that consists of two recurrent neural networks (RNN). One RNN encodes a sequence of symbols into a fixed-length vector representation, and the other decodes the representation into another sequence of symbols. The encoder and decoder of the proposed model are jointly trained to maximize the conditional probability of a target sequence given a source sequence. The performance of a statistical machine translation system is empirically found to improve by using the conditional probabilities of phrase pairs computed by the RNN Encoder-Decoder as an additional feature in the existing log-linear model. Qualitatively, we show that the proposed model learns a semantically and syntactically meaningful representation of linguistic phrases.
Description
Learning Phrase Representations using RNN Encoder-Decoder for
Statistical Machine Translation
%0 Journal Article
%1 Cho2014a
%A Cho, Kyunghyun
%A van Merrienboer, Bart
%A Gulcehre, Caglar
%A Bahdanau, Dzmitry
%A Bougares, Fethi
%A Schwenk, Holger
%A Bengio, Yoshua
%D 2014
%K final thema:seqtoseq
%T Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation
%U http://arxiv.org/abs/1406.1078
%X In this paper, we propose a novel neural network model called RNN Encoder-Decoder that consists of two recurrent neural networks (RNN). One RNN encodes a sequence of symbols into a fixed-length vector representation, and the other decodes the representation into another sequence of symbols. The encoder and decoder of the proposed model are jointly trained to maximize the conditional probability of a target sequence given a source sequence. The performance of a statistical machine translation system is empirically found to improve by using the conditional probabilities of phrase pairs computed by the RNN Encoder-Decoder as an additional feature in the existing log-linear model. Qualitatively, we show that the proposed model learns a semantically and syntactically meaningful representation of linguistic phrases.
@article{Cho2014a,
abstract = {In this paper, we propose a novel neural network model called RNN Encoder-Decoder that consists of two recurrent neural networks (RNN). One RNN encodes a sequence of symbols into a fixed-length vector representation, and the other decodes the representation into another sequence of symbols. The encoder and decoder of the proposed model are jointly trained to maximize the conditional probability of a target sequence given a source sequence. The performance of a statistical machine translation system is empirically found to improve by using the conditional probabilities of phrase pairs computed by the RNN Encoder-Decoder as an additional feature in the existing log-linear model. Qualitatively, we show that the proposed model learns a semantically and syntactically meaningful representation of linguistic phrases.},
added-at = {2018-11-26T22:24:52.000+0100},
archiveprefix = {arXiv},
arxivid = {1406.1078},
author = {Cho, Kyunghyun and van Merrienboer, Bart and Gulcehre, Caglar and Bahdanau, Dzmitry and Bougares, Fethi and Schwenk, Holger and Bengio, Yoshua},
biburl = {https://www.bibsonomy.org/bibtex/2f76f3d93dc9aa690a7e3081c3414cdc9/habereder},
description = {Learning Phrase Representations using RNN Encoder-Decoder for
Statistical Machine Translation},
eprint = {1406.1078},
file = {:C$\backslash$:/Users/Usuario/AppData/Local/Mendeley Ltd./Mendeley Desktop/Downloaded/Cho et al. - 2014 - Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation.pdf:pdf},
interhash = {a4bf56db9d1f80d8681c1b47de0569b3},
intrahash = {f76f3d93dc9aa690a7e3081c3414cdc9},
keywords = {final thema:seqtoseq},
mendeley-tags = {MTM16,ciencia{\_}computacional,traducci{\'{o}}n{\_}autom{\'{a}}tica{\_}neuronal},
month = jun,
timestamp = {2018-11-26T22:24:52.000+0100},
title = {Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translation},
url = {http://arxiv.org/abs/1406.1078},
year = 2014
}