File Download
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.18653/v1/D17-1210
- Scopus: eid_2-s2.0-85054212304
Supplementary
-
Citations:
- Scopus: 0
- Appears in Collections:
Conference Paper: Trainable greedy decoding for neural machine translation
Title | Trainable greedy decoding for neural machine translation |
---|---|
Authors | |
Issue Date | 2017 |
Publisher | Association for Computational Linguistics. The Proceedings' web site is located at http://aclweb.org/anthology/D/D17/#1000 |
Citation | Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), Copenhagen, Denmark, 9-11 September 2017, p. 1968–1978 How to Cite? |
Abstract | Recent research in neural machine translation has largely focused on two aspects; neural network architectures and end-toend learning algorithms. The problem of decoding, however, has received relatively little attention from the research community. In this paper, we solely focus on the problem of decoding given a trained neural machine translation model. Instead of trying to build a new decoding algorithm for any specific decoding objective, we propose the idea of trainable decoding algorithm
in which we train a decoding algorithm to find a translation that maximizes an arbitrary decoding objective. More specifically, we design an actor that observes and manipulates the hidden state
of the neural machine translation decoder and propose to train it using a variant of deterministic policy gradient. We extensively evaluate the proposed algorithm using four language pairs and two decoding
objectives, and show that we can indeed train a trainable greedy decoder that generates a better translation (in terms of a target decoding objective) with minimal computational overhead. |
Persistent Identifier | http://hdl.handle.net/10722/262431 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Gu, J | - |
dc.contributor.author | Cho, K | - |
dc.contributor.author | Li, VOK | - |
dc.date.accessioned | 2018-09-28T04:59:13Z | - |
dc.date.available | 2018-09-28T04:59:13Z | - |
dc.date.issued | 2017 | - |
dc.identifier.citation | Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), Copenhagen, Denmark, 9-11 September 2017, p. 1968–1978 | - |
dc.identifier.uri | http://hdl.handle.net/10722/262431 | - |
dc.description.abstract | Recent research in neural machine translation has largely focused on two aspects; neural network architectures and end-toend learning algorithms. The problem of decoding, however, has received relatively little attention from the research community. In this paper, we solely focus on the problem of decoding given a trained neural machine translation model. Instead of trying to build a new decoding algorithm for any specific decoding objective, we propose the idea of trainable decoding algorithm in which we train a decoding algorithm to find a translation that maximizes an arbitrary decoding objective. More specifically, we design an actor that observes and manipulates the hidden state of the neural machine translation decoder and propose to train it using a variant of deterministic policy gradient. We extensively evaluate the proposed algorithm using four language pairs and two decoding objectives, and show that we can indeed train a trainable greedy decoder that generates a better translation (in terms of a target decoding objective) with minimal computational overhead. | - |
dc.language | eng | - |
dc.publisher | Association for Computational Linguistics. The Proceedings' web site is located at http://aclweb.org/anthology/D/D17/#1000 | - |
dc.relation.ispartof | Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP) | - |
dc.rights | This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. | - |
dc.title | Trainable greedy decoding for neural machine translation | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Li, VOK: vli@eee.hku.hk | - |
dc.identifier.authority | Li, VOK=rp00150 | - |
dc.description.nature | published_or_final_version | - |
dc.identifier.doi | 10.18653/v1/D17-1210 | - |
dc.identifier.scopus | eid_2-s2.0-85054212304 | - |
dc.identifier.hkuros | 292192 | - |
dc.identifier.spage | 1968 | - |
dc.identifier.epage | 1978 | - |