Improving Phrase-Based Statistical Translation Through Combination of Word Alignments

2006 ◽  
pp. 356-367 ◽  
Author(s):  
Boxing Chen ◽  
Marcello Federico
Keyword(s):  
Word Alignments

scholarly journals Unsupervised Sub-tree Alignment for Tree-to-Tree Translation

2013 ◽  
Vol 48 ◽  
pp. 733-782 ◽  
Author(s):  
T. Xiao ◽  
J. Zhu
Keyword(s):  
Machine Translation ◽  
State Of The Art ◽  
Experimental Results ◽  
Alignment Accuracy ◽  
Syntactic Structures ◽  
Tree Alignment ◽  
Matrix Encoding ◽  
Word Alignments ◽  
Alignment Model ◽  
Translation Systems

This article presents a probabilistic sub-tree alignment model and its application to tree-to-tree machine translation. Unlike previous work, we do not resort to surface heuristics or expensive annotated data, but instead derive an unsupervised model to infer the syntactic correspondence between two languages. More importantly, the developed model is syntactically-motivated and does not rely on word alignments. As a by-product, our model outputs a sub-tree alignment matrix encoding a large number of diverse alignments between syntactic structures, from which machine translation systems can efficiently extract translation rules that are often filtered out due to the errors in 1-best alignment. Experimental results show that the proposed approach outperforms three state-of-the-art baseline approaches in both alignment accuracy and grammar quality. When applied to machine translation, our approach yields a +1.0 BLEU improvement and a -0.9 TER reduction on the NIST machine translation evaluation corpora. With tree binarization and fuzzy decoding, it even outperforms a state-of-the-art hierarchical phrase-based system.

scholarly journals Learning Tractable Word Alignment Models with Complex Constraints

2010 ◽  
Vol 36 (3) ◽  
pp. 481-504 ◽  
Author(s):  
João V. Graça ◽  
Kuzman Ganchev ◽  
Ben Taskar
Keyword(s):  
Probabilistic Models ◽  
Learning Algorithm ◽  
Word Alignment ◽  
Word Level ◽  
Word Alignments ◽  
Symmetry Constraints ◽  
Critical Resource ◽  
Complex Constraints ◽  
Bilingual Text ◽  
Efficient Learning

Word-level alignment of bilingual text is a critical resource for a growing variety of tasks. Probabilistic models for word alignment present a fundamental trade-off between richness of captured constraints and correlations versus efficiency and tractability of inference. In this article, we use the Posterior Regularization framework (Graça, Ganchev, and Taskar 2007) to incorporate complex constraints into probabilistic models during learning without changing the efficiency of the underlying model. We focus on the simple and tractable hidden Markov model, and present an efficient learning algorithm for incorporating approximate bijectivity and symmetry constraints. Models estimated with these constraints produce a significant boost in performance as measured by both precision and recall of manually annotated alignments for six language pairs. We also report experiments on two different tasks where word alignments are required: phrase-based machine translation and syntax transfer, and show promising improvements over standard methods.

scholarly journals A Systematic Comparison of Various Statistical Alignment Models

2003 ◽  
Vol 29 (1) ◽  
pp. 19-51 ◽  
Author(s):  
Franz Josef Och ◽  
Hermann Ney
Keyword(s):  
Training Algorithm ◽  
Dice Coefficient ◽  
Design Decisions ◽  
First Order ◽  
Model Yield ◽  
Alignment System ◽  
Word Alignments ◽  
Alignment Model ◽  
Statistical Alignment

We present and compare various methods for computing word alignments using statistical or heuristic models. We consider the five alignment models presented in Brown, Della Pietra, Della Pietra, and Mercer (1993), the hidden Markov alignment model, smoothing techniques, and refinements. These statistical models are compared with two heuristic models based on the Dice coefficient. We present different methods for combining word alignments to perform a symmetrization of directed statistical alignment models. As evaluation criterion, we use the quality of the resulting Viterbi alignment compared to a manually produced reference alignment. We evaluate the models on the German-English Verbmobil task and the French-English Hansards task. We perform a detailed analysis of various design decisions of our statistical alignment system and evaluate these on training corpora of various sizes. An important result is that refined alignment models with a first-order dependence and a fertility model yield significantly better results than simple heuristic models. In the Appendix, we present an efficient training algorithm for the alignment models presented.

Constructing Corpora for the Development and Evaluation of Paraphrase Systems

2008 ◽  
Vol 34 (4) ◽  
pp. 597-614 ◽  
Author(s):  
Trevor Cohn ◽  
Chris Callison-Burch ◽  
Mirella Lapata
Keyword(s):  
Natural Language Processing ◽  
Natural Language ◽  
Language Processing ◽  
Syntactic Structure ◽  
Nominal Data ◽  
Parallel Corpus ◽  
Word Alignments ◽  
Definition Of

Automatic paraphrasing is an important component in many natural language processing tasks. In this article we present a new parallel corpus with paraphrase annotations. We adopt a definition of paraphrase based on word alignments and show that it yields high inter-annotator agreement. As Kappa is suited to nominal data, we employ an alternative agreement statistic which is appropriate for structured alignment tasks. We discuss how the corpus can be usefully employed in evaluating paraphrase systems automatically (e.g., by measuring precision, recall, and F1) and also in developing linguistically rich paraphrase models based on syntactic structure.

scholarly journals Visualization, Search and Analysis of Hierarchical Translation Equivalence in Machine Translation Data

2014 ◽  
Vol 101 (1) ◽  
pp. 43-54 ◽  
Author(s):  
Gideon Maillette de Buy Wenniger ◽  
Khalil Sima’an
Keyword(s):  
Machine Translation ◽  
Equivalence Relations ◽  
Qualitative And Quantitative Analysis ◽  
Complete Representation ◽  
Qualitative And Quantitative ◽  
Hierarchical Relations ◽  
Word Alignments ◽  
Detailed Statistical Analysis ◽  
Translation Systems ◽  
Search Capability

Abstract Translation equivalence constitutes the basis of all Machine Translation systems including the recent hierarchical and syntax-based systems. For hierarchical MT research it is important to have a tool that supports the qualitative and quantitative analysis of hierarchical translation equivalence relations extracted from word alignments in data. In this paper we present such a toolkit and exemplify some of its uses. The main challenges taken up in designing this tool are the efficient and compact, yet complete, representation of hierarchical translation equivalence coupled with an intuitive visualization of these hierarchical relations. We exploit a new hierarchical representation, called Hierarchical Alignment Trees (HATs), which is based on an extension of the algorithms used for factorizing n-ary branching SCFG rules into their minimally-branching equivalents. Our toolkit further provides a search capability based on hierarchically relevant properties of word alignments and/or translation equivalence relations. Finally, the tool allows detailed statistical analysis of word alignments, thereby providing a breakdown of alignment statistics according to the complexity of translation equivalence units or reordering phenomena. We illustrate this with an empirical study of the coverage of inversion-transduction grammars for a number of corpora enriched with manual or automatic word alignments, followed by a breakdown of corpus statistics to reordering complexity.

Sign in / Sign up

Export Citation Format

Share Document