Experiments in Domain Adaptation for Statistical Machine Translation

By Philipp Koehn, Josh Schroeder. WMT@ACL, 2007

360 citations recorded by the source corpus.

The special challenge of the WMT 2007 shared task was domain adaptation. We took this opportunity to experiment with various ways of adapting a statistical machine translation systems to a special domain (here: news commentary), when most of the training data is from a different domain (here: European Parliament speeches). This paper also gives a description of the submission of the University of Edinburgh to the shared task.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/98d98952e84d3352398dab9eb01de4c17fd0f0c6

How researchers see this paper

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

In Artificial Intelligence

Citations in the corpus

Freezing Subnetworks to Analyze Domain Adaptation in Neural Machine Translation (2018, background): “, 2011), or on interpolating in-domain and out-of-domain models (Lu et al., 2007; Foster et al., 2010; Koehn and Schroeder, 2007).”
Language Models for Machine Translation: Original vs. Translated Texts (2011, methodology): “The feature weight of each LM is set by minimum error-rate tuning, optimizing the translation quality; this is the same technique that Koehn and Schroeder (2007) employ for domain adaptation.”
PPDB: The Paraphrase Database (2013, background): “, 2009), the Czech, German, Spanish and French portions of the News Commentary data (Koehn and Schroeder, 2007), the United Nations French- and Spanish-English parallel corpora (Eisele and Chen, 2010), the JRC Acquis corpus (Steinberger et al.”
Counterfactual Learning from Bandit Feedback under Deterministic Logging : A Case Study in Statistical Machine Translation (2017, methodology): “We conduct two SMT tasks with hypergraph re-decoding: The first is German-to-English and is trained using a concatenation of the Europarl corpus (Koehn, 2005), the Common Crawl corpus3 and the News Commentary corpus (Koehn and Schroeder, 2007).”
Combining Domain Adaptation Approaches for Medical Text Translation (2014, methodology): “The additional model is log-linearly interpolated with the indomain model (Baseline2) using the multidecoding method described in (Koehn and Schroeder, 2007).”
Fast Domain Adaptation for Neural Machine Translation (2016, methodology): “Second, different methods of interpolating in-domain and out-of-domain models [L¨u et al., 2007, Koehn and Schroeder, 2007, Foster and Kuhn, 2007] have been proposed.”
Finding Sparse Structures for Domain Specific Neural Machine Translation (2021, background): “It is also observed that NMT does not perform well in specific domains where the domain-specific corpora are limited, such as medical domain (Koehn and Schroeder 2007; Axelrod, He, and Gao 2011; Freitag and Al-Onaizan 2016; Chu and Wang 2018).”
Domain Adaptation via Pseudo In-Domain Data Selection (2011, methodology): “Lastly, Koehn and Schroeder (2007) reported improvements from using multiple decoding paths (Birch et al., 2007) to pass both tables to the Moses SMT decoder (Koehn et al., 2003), instead of directly combining the phrase tables to perform domain adaptation.”
Domain Adaptation for Medical Text Translation using Web Resources (2014, methodology): “For TM, the additional model is log-linearly interpolated with the in-domain model using the multi-decoding method described in (Koehn and Schroeder, 2007).”
Steering Output Style and Topic in Neural Response Generation (2017, methodology): “Similar ideas have been tested in domain adaptation for statistical machine translation (Koehn and Schroeder, 2007), where both in-domain and open-domain translation tables were used as candidates for generating target sentence.”
Assessing the Tolerance of Neural Machine Translation Systems Against Speech Recognition Errors (2017, methodology): “Our phrase-based MT system (which we refer to as MMT) is built upon the ModernMT framework: an extension of the phrase-based MT framework introduced in [16] that enables context-aware translation for fast and incremental training.”
Domain Adaptation for Statistical Machine Translation with Monolingual Resources (2009, methodology): “Our work is mostly related to (Koehn and Schroeder, 2007) but explores different assumptions about available adaptation data: i.e. only monolingual in-domain texts are available.”
Bi-Directional Neural Machine Translation with Synthetic Parallel Data (2018, background): “SMT systems have the advantage of a dedicated language model that can incorporate all available target-side monolingual data to signiﬁ-cantly improve translation quality (Koehn et al., 2003; Koehn and Schroeder, 2007).”
Adaptation of machine translation for multilingual information retrieval in the medical domain (2014, methodology): “To make this information explicitly available, we follow the approach of Koehn and Schroeder [20] and train two independent translation models (phrase tables).”
Factored Translation Models (2007, methodology): “Factored translation models have also been used for the integration of CCG supertags (Birch et al., 2007), domain adaptation (Koehn and Schroeder, 2007) and for the improvement of English-Czech translation (Bojar, 2007).”
Multi-domain Neural Network Language Generation for Spoken Dialogue Systems (2016, methodology): “Discriminative training (DT) has been successfully used to train RNNs for various tasks.”
Multi-Domain Neural Machine Translation through Unsupervised Adaptation (2017, methodology): “This dataset is then used to train speciﬁc models that are interpolated with generic ones using standard log-linear methods (Koehn and Schroeder, 2007) or mixture models (Foster and Kuhn, 2007).”