Sense-Aware Statistical Machine Translation using Adaptive Context-Dependent Clustering

Proceedings of the Conference on Machine Translation (WMT), Volume 1: Research Papers, pages 1–10 Copenhagen, Denmark, September 711, 2017. c⃝2017 Association for Computational Linguistics Sense-Aware Statistical Machine Translation using Adaptive Context-Dependent Clustering Xiao Pu EPFL & Idiap Research Inst. Martigny, Switzerland xiao.pu@idiap.ch Nikolaos Pappas Idiap Research Institute Martigny, Switzerland nikolaos.pappas@idiap.ch Andrei Popescu-Belis Idiap Research Institute Martigny, Switzerland apbelis@idiap.ch Abstract Statistical machine translation (SMT) sys- tems use local cues from n-gram trans- lation and language models to select the translation of each source word. Such systems do not explicitly perform word sense disambiguation (WSD), although this would enable them to select transla- tions depending on the hypothesized sense of each word. Previous attempts to con- strain word translations based on the re- sults of generic WSD systems have suf- fered from their limited accuracy. We demonstrate that WSD systems can be adapted to help SMT, thanks to three key achievements: (1) we consider a larger context for WSD than SMT can afford to consider; (2) we adapt the number of senses per word to the ones observed in the training data using clustering-based WSD with K-means; and (3) we initialize sense- clustering with definitions or examples ex- tracted from WordNet. Our WSD system is competitive, and in combination with a factored SMT system improves noun and verb translation from English to Chinese, Dutch, French, German, and Spanish. 1 Introduction Selecting the correct translation of polysemous words remains an important challenge for ma- chine translation (MT). While some translation options may be interchangeable, substantially dif- ferent senses of source words must generally be rendered by different words in the target language. In this case, an MT system should identify – im- plicitly or explicitly – the correct sense conveyed by each occurrence in order to select the appropri- ate translation. Source: And I do really like this shot, be- cause it shows all the detritus that’s sort of embedded in the sole of the sneakers. Baseline SMT: Und ich mag dieses Bild ... Online NMT: Und ich mag diesen Schuss wirklich, ... Sense-aware MT: Und ich mag diese Auf- nahme wirklich, ... Reference translation: Ich mag diese Auf- nahme wirklich, ... Figure 1: Example of sense-aware translation that is closer to a reference translation than a baseline statistical MT system or an online neural one. Current statistical or neural MT systems per- form word sense disambiguation (WSD) implic- itly, for instance through the n-gram frequency in- formation stored in the translation and language models. However, the context taken into account by an MT system when performing implicit WSD is limited. For instance, in the case of phrase- based SMT, it is the order of the language model (often between 3 and 5) and the length of n-grams in the phrase table (seldom above 5). In attention- based neural MT systems, the context extends to the entire sentence, but is not specifically trained to be used for WSD. For instance, Figure 1 shows an English sen- tence translated into German by a baseline …

Content truncated to 500 words