Edinburgh’s Machine Translation Systems for European Language Pairs

By Nadir Durrani, Barry Haddow, Kenneth Heafield, Philipp Koehn. 2013

43 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): “For the English-to-German sys-tem we added an OSM model over [pos, morph] (source:pos, target:morph) and for the German-to-English system we added an OSM model over [morph,pos] (source:morph, target:pos), a conﬁg-uration that was found to work best in our previous experiments (Birch et al., 2013).”
EU-BRIDGE MT: Combined Machine Translation (2014, methodology): “UEDIN’s phrase-based systems (Durrani et al., 2014) have been trained using the Moses toolkit (Koehn et al., 2007), replicating the settings described in (Durrani et al., 2013b).”
Integrating an Unsupervised Transliteration Model into Statistical Machine Translation (2014, methodology): “See Table 1 Baseline Settings: We trained a Moses system replicating the settings used in competition-grade systems (Durrani et al., 2013b; Birch et al., 2013): a maximum sentence length of 80, GDFA sym-metrization of GIZA++ alignments (Och and Ney, 2003), an interpolated Kneser-Ney smoothed 5-gram…”
English SLT and MT system description for the IWSLT 2013 evaluation (2013, methodology): “[41] N. Durrani, H. Sajjad, H. Hoang, and P. Koehn, “Integrating an Unsupervised Transliteration Model into Statistical Machine Translation,” in Proceedings of the 15th Conference of the European Chapter of the ACL (EACL 2014).”
Findings of the 2013 Workshop on Statistical Machine Translation (2013, unknown): “UEDIN, UEDIN-HEAFIELD University of Edinburgh (Durrani et al., 2013b)”
Phrasal: A Toolkit for New Directions in Statistical Machine Translation (2014, background): “Phrasal can load multiple language models, and includes native support for the class-based language models that have become popular in recent evaluations (Wuebker et al., 2012; Ammar et al., 2013; Durrani et al., 2013).”
Scalable Modified Kneser-Ney Language Model Estimation (2013, background): “, 2007) submission (Durrani et al., 2013), which used all constrained data specified by the evaluation (7 billion tokens of English).”
Target-Centric Features for Translation Quality Estimation (2014, methodology): “We extract features from the following components of the Moses de-coder: distortion model, language model, lexical reordering, lexical translation probability, operational sequence model (Durrani et al., 2013), phrase translation probability, and the decoder score.”
Yandex School of Data Analysis Russian-English Machine Translation System for WMT14 (2014, unknown): “Please, refer to (Durrani et al., 2013) for implementation details.”
Manawi: Using Multi-Word Expressions and Named Entities to Improve Machine Translation (2014, methodology): “For instance, the Moore-Lewis ﬁlter re-moves sentence pairs based on source-side cross-entropy differences (Moore and Lewis, 2010) and the Edinburgh’s MT system used the Modiﬁed Moore-Lewis ﬁltering (Axelrod et al., 2011) in WMT 2013 shared task (Durrani et al., 2013).”
Translation Task : Two-step Data Selection and RBMT-Style Synthetic Rules (2014, methodology): “Finally, we make use of the findings from WMT 2013 brought by the winning team (Durrani et al., 2013) and decide to use the Operation Sequence Model (OSM), based on minimal translation units and Markov chains over sequences of operations, implemented in MOSES 1LDC2011T07 English Gigaword Fifth…”