Experiments in domain adaptation for statistical machine translation

Koehn, Philipp; Schroeder, Josh E.

doi:10.3115/1626355.1626388

Cited by 209 publications

(158 citation statements)

References 7 publications

Supporting

Mentioning

153

Contrasting

Unclassified

Order By: Relevance

“…Some authors attempt to combine the predictions of two separate (in-domain and general-domain) translation models [16][17][18][19] or language models [20]. Wu and Wang [21] use in-domain data to improve word alignment in the training phase.…”

Section: Domain Adaptationmentioning

confidence: 99%

“…To make this information explicitly available, we follow the approach of Koehn and Schroeder [20] and train two independent translation models (phrase tables). The first experiment is based on the P5 configuration of parallel data: one phrase table is trained using in-domain sections of the training data (dictionaries and medical corpora) and the second using data selected from the general domain.…”

Section: Optimization Of Phrase Table Configurationmentioning

confidence: 99%

See 1 more Smart Citation

Adaptation of machine translation for multilingual information retrieval in the medical domain

Pecina¹,

Dušek²,

Goeuriot³

et al. 2014

Artificial Intelligence in Medicine

View full text Add to dashboard Cite

Objective. We investigate machine translation (MT) of user search queries in the context of cross-lingual information retrieval (IR) in the medical domain. The main focus is on techniques to adapt MT to increase translation quality; however, we also explore MT adaptation to improve effectiveness of cross-lingual IR.Methods and Data. Our MT system is Moses, a state-of-the-art phrase-based statistical machine translation system. The IR system is based on the BM25 retrieval model implemented in the Lucene search engine. The MT techniques employed in this work include in-domain training and tuning, intelligent training data selection, optimization of phrase table configuration, compound splitting, and exploiting synonyms as translation variants. The IR methods include morphological normalization and using multiple translation variants for query expansion. The experiments are performed and thoroughly evaluated on three language pairs: Czech-English, German-English, and French-English. MT quality is evaluated on data sets created within the Khresmoi project and IR effectiveness is tested on the CLEF eHealth 2013 data sets.Results. The search query translation results achieved in our experiments are outstanding -our systems outperform not only our strong baselines, but also Google Translate and Microsoft Bing Translator in direct comparison carried out on all the language pairs. The baseline BLEU scores increased from 26.59 to 41.45 for Czech-English, from 23.03 to 40.82 for German-English, and from 32.67 to 40.82 for French-English. This is a 55% improvement on average. In terms of the IR performance on this particular test collection, a significant improvement over the baseline is achieved only for French-English. For Czech-English and German-English, the increased MT quality does not lead to better IR results.Conclusions. Most of the MT techniques employed in our experiments improve MT of medical search queries. Especially the intelligent training data selection proves to be very successful for domain adaptation of MT. Certain improvements are also obtained from German compound splitting on the source language side. Translation quality, however, does not appear to correlate with the IR performance -better translation does not necessarily yield better retrieval. We discuss in detail the contribution of the individual techniques and state-of-the-art features and provide future research directions.

show abstract

Section: Domain Adaptationmentioning

confidence: 99%

Section: Optimization Of Phrase Table Configurationmentioning

confidence: 99%

Adaptation of machine translation for multilingual information retrieval in the medical domain

Pecina¹,

Dušek²,

Goeuriot³

et al. 2014

Artificial Intelligence in Medicine

View full text Add to dashboard Cite

show abstract

“…Other authors [24,3] studied different ways to combine bilingual or only source data from different domains. The use of clustering to extract sub-domains to build more specific language or translation models has also been studied in [54,43].…”

Section: Adaptationmentioning

confidence: 99%

Online Learning via Dynamic Reranking for Computer Assisted Translation

Martínez-Gómez

Sanchis-Trilles

Casacuberta

2011

Computational Linguistics and Intelligent Text Processing

View full text Add to dashboard Cite

“…Following the ideas in [1], one of the first works was performed in [7], where the authors added cache language and translation models to an interactive machine translation system. In [3], different ways to combine the available data belonging to two different sources were studied. The work in [8] explores alignment model mixtures as a way of performing topic adaptation.…”

Section: Related Workmentioning

confidence: 99%

“…Adaptation has become a very popular issue in natural language processing [1,2,3], and more specifically in statistical machine translation (SMT) [4]. Typically, the adaptation problem arises when two very different sets of training data are available, yielding two different sets of model parameters.…”

Section: Introductionmentioning

confidence: 99%

Improving translation quality stability using Bayesian predictive adaptation

Sanchis-Trilles

Casacuberta

2015

Computer Speech & Language

View full text Add to dashboard Cite

We introduce a Bayesian approach for the adaptation of the log-linear weights present in state-of-the-art statistical machine translation systems. Typically, these weights are estimated by optimising a given translation quality criterion, taking only into account a certain set of development data (e.g., the adaptation data). In this article, we show that the Bayesian framework provides appropriate estimates of such weights in conditions where adaptation data is scarce. The theoretical framework is presented, alongside with a thorough experimentation and comparison with other weight estimation methods. We provide a comparison of different sampling strategies, including an effective heuristic strategy and a theoretically sound Markov chain Monte-Carlo algorithm. Experimental results show that Bayesian predictive adaptation (BPA) outperforms the reestimation from scratch in conditions where adaptation data is scarce. Further analysis reveals that the improvements obtained are due to the greater stability of the estimation procedure. In addition, the proposed BPA framework has a much lower computational cost than raw re-estimation.

show abstract

Experiments in domain adaptation for statistical machine translation

Cited by 209 publications

References 7 publications

Adaptation of machine translation for multilingual information retrieval in the medical domain

Adaptation of machine translation for multilingual information retrieval in the medical domain

Online Learning via Dynamic Reranking for Computer Assisted Translation

Improving translation quality stability using Bayesian predictive adaptation

Contact Info

Product

Resources

About