Word Sense Disambiguation Improves Statistical Machine Translation

By Yee Seng Chan, Hwee Tou Ng, David Chiang. Annual Meeting of the Association for Computational Linguistics, 2007

328 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/bbe013543e9c1d8f00499036121c363ad3ab3d7d

How researchers see this paper

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

Citations in the corpus

On internal language representations in deep learning: an analysis of machine translation and speech recognition (2018, methodology): “Super-word levels texts utterances sentences phrases relations words morphemes syllables characters phonemes Sub-word levels systems, for instance by using word sense disambiguation [52, 58, 341].”
It Makes Sense: A Wide-Coverage Word Sense Disambiguation System for Free Text (2010, methodology): “As a fundamental task in natural language processing (NLP), WSD can benefit applications such as machine translation (Chan et al., 2007a; Carpuat and Wu, 2007) and information retrieval (Stokoe et al.”
11,001 New Features for Statistical Machine Translation (2009, methodology): “…Watanabe et al. (2007), but more in the spirit of features used in the word sense disambiguation model introduced by Lee and Ng (2002) and incorporated as a submodel of a translation system by Chan et al. (2007); here, we are incorporating some of its features directly into the translation model.”
Vector Space Models of Lexical Meaning (2015, background): “There is also a long ongoing debate about whether word sense disambiguation is useful more generally for language processing applications (e.g. Chan et al. (2007)).”
The Hitchhiker’s Guide to Testing Statistical Significance in Natural Language Processing (2018, background): “(Collins et al., 2005; Chan et al., 2007; Rush et al., 2012)).”
On the Evaluation of Semantic Phenomena in Neural Machine Translation Using Natural Language Inference (2018, background): “McCann et al. (2017) show improvements in many tasks by using contextualized word vectors extracted from a LSTM encoder trained for MT.”