An End-to-End Discriminative Approach to Machine Translation
作者： Percy Liang , Alexandre Bouchard-Côté , Dan Klein , Ben Taskar
关键词:
摘要: We present a perceptron-style discriminative approach to machine translation in which large feature sets can be exploited. Unlike reranking approaches, our system take advantage of learned features all stages decoding. first discuss several challenges error-driven approaches. In particular, we explore different ways updating parameters given training example. find that making frequent but smaller updates is preferable fewer larger updates. Then, an array and show both how they quantitatively increase BLEU score qualitatively interact on specific examples. One particular investigate novel way introduce learning into the initial phrase extraction process, has previously been entirely heuristic.
core.ac.uk
aclweb.org
acm.org
stanford.edu 参考文章(17)
Dragomir R. Radev, Sanjeev Khudanpur, Daniel Gildea, Katherine Eng, Alexander M. Fraser, Shankar Kumar, Anoop Sarkar, Zhen Jin, Libin Shen, Franz Josef Och, Kenji Yamada, David Smith, Viren Jain, A Smorgasbord of Features for Statistical Machine Translation north american chapter of the association for computational linguistics. pp. 161- 168 ,(2004)
Richard Zens, Hermann Ney, Improvements in Phrase-Based Statistical Machine Translation north american chapter of the association for computational linguistics. pp. 257- 264 ,(2004)
Anoop Sarkar, Libin Shen, Franz Josef Och, Discriminative Reranking for Machine Translation north american chapter of the association for computational linguistics. pp. 177- 184 ,(2004)
Andreas Stolcke, SRILM – An Extensible Language Modeling Toolkit conference of the international speech communication association. ,(2002)
Vincent J. Della Pietra, Stephen A. Della Pietra, Robert L. Mercer, Peter F. Brown, The mathematics of statistical machine translation: parameter estimation Computational Linguistics. ,vol. 19, pp. 263- 311 ,(1993)
Michael Collins, Discriminative training methods for hidden Markov models: theory and experiments with perceptron algorithms empirical methods in natural language processing. pp. 1- 8 ,(2002) , 10.3115/1118693.1118694
Terry Koo, Michael Collins, Hidden-variable models for discriminative reranking Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing - HLT '05. pp. 507- 514 ,(2005) , 10.3115/1220575.1220639
Michael Collins, Brian Roark, Incremental Parsing with the Perceptron Algorithm meeting of the association for computational linguistics. pp. 111- 118 ,(2004) , 10.3115/1218955.1218970
Vincent J. Della Pietra, Jenifer C. Lai, Robert L. Mercer, Peter F. Brown, Peter V. deSouza, Class-based n -gram models of natural language Computational Linguistics. ,vol. 18, pp. 467- 479 ,(1992) , 10.5555/176313.176316
Franz Josef Och, Minimum error rate training in statistical machine translation Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - ACL '03. pp. 160- 167 ,(2003) , 10.3115/1075096.1075117