Don't until the final verb wait: Reinforcement learning for simultaneous machine translation

Alvin C. Grissom; Jordan Boyd-Graber; He He; John Morgan; Hal Daumé

Don't until the final verb wait: Reinforcement learning for simultaneous machine translation

Alvin C. Grissom, Jordan Boyd-Graber, He He, John Morgan, Hal Daumé

Computer Science

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

Abstract

We introduce a reinforcement learningbased approach to simultaneous machine translation-producing a translation while receiving input words- between languages with drastically different word orders: from verb-final languages (e.g., German) to verb-medial languages (English). In traditional machine translation, a translator must "wait" for source material to appear before translation begins. We remove this bottleneck by predicting the final verb in advance. We use reinforcement learning to learn when to trust predictions about unseen, future portions of the sentence. We also introduce an evaluation metric to measure expeditiousness and quality. We show that our new translation model outperforms batch and monotone translation strategies.

Original language	English (US)
Title of host publication	EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference
Publisher	Association for Computational Linguistics (ACL)
Pages	1342-1352
Number of pages	11
ISBN (Electronic)	9781937284961
State	Published - 2014
Event	2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014 - Doha, Qatar Duration: Oct 25 2014 → Oct 29 2014

Publication series

Name	EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

Conference

Conference	2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014
Country/Territory	Qatar
City	Doha
Period	10/25/14 → 10/29/14

ASJC Scopus subject areas

Computational Theory and Mathematics
Computer Vision and Pattern Recognition
Information Systems

Cite this

Grissom, A. C., Boyd-Graber, J., He, H., Morgan, J., & Daumé, H. (2014). Don't until the final verb wait: Reinforcement learning for simultaneous machine translation. In EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference (pp. 1342-1352). (EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference). Association for Computational Linguistics (ACL).

Don't until the final verb wait: Reinforcement learning for simultaneous machine translation. / Grissom, Alvin C.; Boyd-Graber, Jordan; He, He et al.
EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2014. p. 1342-1352 (EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference).

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

Grissom, AC, Boyd-Graber, J, He, H, Morgan, J & Daumé, H 2014, Don't until the final verb wait: Reinforcement learning for simultaneous machine translation. in EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference, Association for Computational Linguistics (ACL), pp. 1342-1352, 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014, Doha, Qatar, 10/25/14.

Grissom AC, Boyd-Graber J, He H, Morgan J, Daumé H. Don't until the final verb wait: Reinforcement learning for simultaneous machine translation. In EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. Association for Computational Linguistics (ACL). 2014. p. 1342-1352. (EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference).

Grissom, Alvin C. ; Boyd-Graber, Jordan ; He, He et al. / Don't until the final verb wait : Reinforcement learning for simultaneous machine translation. EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2014. pp. 1342-1352 (EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference).

@inproceedings{df8d692cf59f4e518906cf917eb6279d,

title = "Don't until the final verb wait: Reinforcement learning for simultaneous machine translation",

abstract = "We introduce a reinforcement learningbased approach to simultaneous machine translation-producing a translation while receiving input words- between languages with drastically different word orders: from verb-final languages (e.g., German) to verb-medial languages (English). In traditional machine translation, a translator must {"}wait{"} for source material to appear before translation begins. We remove this bottleneck by predicting the final verb in advance. We use reinforcement learning to learn when to trust predictions about unseen, future portions of the sentence. We also introduce an evaluation metric to measure expeditiousness and quality. We show that our new translation model outperforms batch and monotone translation strategies.",

author = "Grissom, {Alvin C.} and Jordan Boyd-Graber and He He and John Morgan and Hal Daum{\'e}",

note = "Publisher Copyright: {\textcopyright} 2014 Association for Computational Linguistics.; 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014 ; Conference date: 25-10-2014 Through 29-10-2014",

year = "2014",

language = "English (US)",

series = "EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference",

publisher = "Association for Computational Linguistics (ACL)",

pages = "1342--1352",

booktitle = "EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference",

}

TY - GEN

T1 - Don't until the final verb wait

T2 - 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014

AU - Grissom, Alvin C.

AU - Boyd-Graber, Jordan

AU - He, He

AU - Morgan, John

AU - Daumé, Hal

PY - 2014

Y1 - 2014

N2 - We introduce a reinforcement learningbased approach to simultaneous machine translation-producing a translation while receiving input words- between languages with drastically different word orders: from verb-final languages (e.g., German) to verb-medial languages (English). In traditional machine translation, a translator must "wait" for source material to appear before translation begins. We remove this bottleneck by predicting the final verb in advance. We use reinforcement learning to learn when to trust predictions about unseen, future portions of the sentence. We also introduce an evaluation metric to measure expeditiousness and quality. We show that our new translation model outperforms batch and monotone translation strategies.

AB - We introduce a reinforcement learningbased approach to simultaneous machine translation-producing a translation while receiving input words- between languages with drastically different word orders: from verb-final languages (e.g., German) to verb-medial languages (English). In traditional machine translation, a translator must "wait" for source material to appear before translation begins. We remove this bottleneck by predicting the final verb in advance. We use reinforcement learning to learn when to trust predictions about unseen, future portions of the sentence. We also introduce an evaluation metric to measure expeditiousness and quality. We show that our new translation model outperforms batch and monotone translation strategies.

UR - http://www.scopus.com/inward/record.url?scp=84961288938&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84961288938&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84961288938

T3 - EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

SP - 1342

EP - 1352

BT - EMNLP 2014 - 2014 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

PB - Association for Computational Linguistics (ACL)

Y2 - 25 October 2014 through 29 October 2014

ER -

Don't until the final verb wait: Reinforcement learning for simultaneous machine translation

Abstract

Publication series

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this