A Joint Sequence Translation Model with Integrated Reordering

By Nadir Durrani, Helmut Schmid, Alexander M. Fraser. 2011

134 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

Citations in the corpus

Edinburgh’s Phrase-based Machine Translation Systems for WMT-14 (2014, methodology): “In order to create systems for the medical translation tasks, we used phrase-based Moses with exactly the same settings as for the news translation task, including the OSM (Durrani et al., 2011), and compound splitting Koehn and Knight (2003) for German source.”
Findings of the 2016 Conference on Machine Translation (2016, methodology): “This system combines the Operation Sequence Model (OSM) (Durrani et al., 2011) with the classic phrase-based statistical MT (PB-SMT) approach.”
Paraphrasing Revisited with Neural Machine Translation (2017, methodology): “which has been shown to improve the performance of SMT systems (Durrani et al., 2011).”
Word Representation Models for Morphologically Rich Languages in Neural Machine Translation (2016, methodology): “Our OSM can be considered as a form of attentional encoder-decoder Bahdanau et al. (2015) with hard attention in which each decision is con-textualised by at most one source word, contrasting with the soft attention in Bahdanau et al. (2015).”
The AFRL-MITLL WMT16 News-Translation Task Systems (2016, methodology): “…with the following parameters/features: distortion-limit of 8, no reordering over punctuation, hierarchical mslr reordering model (Galley and Manning, 2008), order 7 operational sequence model (Durrani et al., 2011), and a factored language model over the NYT Gi-gaword corpus with 600 word classes.”
Machine Translation and Monolingual Postediting: The AFRL WMT-14 System (2014, methodology): “1 of moses, and incorporated additional decoding features, namely the Operation Sequence Model (Durrani et al., 2011) and Lexicalized Reordering Model (Tillman, 2004; Galley and Manning, 2008).”
Edinburgh SLT and MT system description for the IWSLT 2014 evaluation (2014, background): “• Phrase translation scores in both directions, smoothed with Good-Turing discounting • Lexical translation scores in both directions • Word and phrase penalties • Six simple count-based binary features • Phrase length features • Distance-based distortion cost • A hierarchical lexicalized reordering”
Translation Task : Two-step Data Selection and RBMT-Style Synthetic Rules (2014, methodology): “…units and Markov chains over sequences of operations, implemented in MOSES 1LDC2011T07 English Gigaword Fifth Edition, LDC2011T10 French Gigaword Third Edition 2This ratio was empirically chosen based on words fertility between English and French. and introduced by Durrani et al. (2011).”
Abu-MaTran at WMT 2015 Translation Task: Morphological Segmentation and Web Crawling (2015, methodology): “The phrase-based SMT system is augmented with additional features: an Operation Sequence Model (OSM) (Durrani et al., 2011) and a Bilingual Neural Language Model (BiNLM) (De-vlin et al., 2014), both trained on the parallel data used to learn the phrase-table.”
Can Markov Models Over Minimal Translation Units Help Phrase-Based SMT? (2013, methodology): “We deﬁne the main OSM feature, and borrow 4 supportive features, the Gap , Open Gap , Gap-width and Deletion penalties (Durrani et al., 2011).”
Model With Minimal Translation Units, But Decode With Phrases (2013, methodology): “Durrani et al. (2011) recently addressed these problems by proposing an operation sequence N-gram model which strongly couples translation and reordering, hypothesizes all possible reorderings and does not require POS-based rules.”
Farasa: A Fast and Furious Segmenter for Arabic (2016, methodology): “Among other features, we used lexical-ized reordering model (Galley and Manning, 2008), a 5-gram Operation Sequence Model (Durrani et al., 2011), Class-based Models (Durrani et al., 2014b) and other default parameters.”
USAAR: An Operation Sequential Model for Automatic Statistical Post-Editing (2016, methodology): “We inspired from the work of Durrani et al. (2011) and Durrani et al. (2015).”
English SLT and MT system description for the IWSLT 2013 evaluation (2013, methodology): “• operation sequence model (OSM) with 4 additional supportive features: 2 gap based penalties, 1 distance based feature and 1 deletion penalty [25]”
A Multifaceted Evaluation of Neural versus Phrase-Based Machine Translation for 9 Language Directions (2017, methodology): “, 2016) NMT Same as for EN→CS EN→FI PBMT Phrase-based, rule-based and unsupervised word segmentation, operation sequence model (Durrani et al., 2011), bilingual neural language model (Devlin et al.”
DCU-Lingo24 Participation in WMT 2014 Hindi-English Translation task (2014, methodology): “, 2012), Operation Sequence Model (OSM) (Durrani et al., 2011) and Language Model Interpolation(LMI).”
Edinburgh’s Machine Translation Systems for European Language Pairs (2013, methodology): “We enhanced the phrase segmentation and re-ordering mechanism by integrating OSM: an operation sequence N-gram-based translation and re-ordering model (Durrani et al., 2011) into the Moses phrase-based decoder.”
Investigating the Usefulness of Generalized Word Representations in SMT (2014, methodology): “The Operation Sequence Model (Durrani et al., 2011) is an instance of the N-gram based SMT framework (Casacuberta and Vidal, 2004; Mari˜no et al., 2006).”