Re-structuring, Re-labeling, and Re-aligning for Syntax-Based Machine Translation
Bibliographic Data
| ID | 12156432 |
|---|---|
| Authors | Wei Wang (0000-0001-5806-2368, 6060 Center Drive, Suite 150, Los Angeles, CA, 90045, USA.., corresponding author), JONATHAN MAY (0000-0002-5284-477X, 6060 Center Drive, Suite 150, Los Angeles, CA, 90045, USA.., corresponding author), Kevin Knight (0000-0001-9117-1718, 6060 Center Drive, Suite 150, Los Angeles, CA, 90045, USA.., corresponding author), Daniel Marcu (6060 Center Drive, Suite 150, Los Angeles, CA, 90045, USA.., corresponding author) |
| Year | 2010 |
| Volume | 36 |
| Issue | 2 |
| Pages | 247-277 |
| Publication date | 2010-03-25 |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | Computational Linguistics (JOURNAL) |
| Journal identifiers | ISSN: 0891-2017 • E-ISSN: 1530-9312 |
| Publisher | Association for Computational Linguistics (PUBLISHER • US) |
| DOI | 10.1162/coli.2010.36.2.09054 |
| OpenAlex | W2167393476 |
| Language | EN |
| Citations received | 1 |
| References cited | 9 |
This article shows that the structure of bilingual material from standard parsing and alignment tools is not optimal for training syntax-based statistical machine translation (SMT) systems. We present three modifications to the MT training data to improve the accuracy of a state-of-the-art syntax MT system: re-structuring changes the syntactic structure of training parse trees to enable reuse of substructures; re-labeling alters bracket labels to enrich rule application context; and re-aligning unifies word alignment across sentences to remove bad word alignments and refine good ones. Better structures, labels, and word alignments are learned by the EM algorithm. We show that each individual technique leads to improvement as measured by BLEU, and we also show that the greatest improvement is achieved by combining them. We report an overall 1.48 BLEU improvement on the NIST08 evaluation set over a strong baseline in Chinese/English translation
Baseline (sea · Context (archaeology · Linguistics · Machine translation · Natural language processing · Parsing · Programming language · Set (abstract data type · Structuring · Syntactic structure · Syntax · Translation (biology · Word (group theory · Computer Science · Natural Language Processing Techniques · Text Readability and Simplification · Topic Modeling · Artificial Intelligence
Maximum Likelihood from Incomplete Data Via the EM Algorithm
Hierarchical Phrase-Based Translation
Modality and Negation in Simt Use of Modality and Negation in Semantically-Informed Syntactic MT
Training Tree Transducers
The Alignment Template Approach to Statistical Machine Translation
The DOP Estimation Method Is Biased and Inconsistent
| Citations per year | 0.07 |
|---|---|
| Citation span | 2012 - 2012 (1) |
| Highly cited | No |