Applying cross-entropy difference for selecting parallel training data from publicly available sources for conversational machine translation

Applying cross-entropy difference for selecting parallel training data from publicly available sources for conversational machine translation William Lewis author Christian Federmann author Ying Xin author 2015-dec 3-4 text Proceedings of the 12th International Workshop on Spoken Language Translation: Papers Marcello Federico editor Sebastian Stüker editor Jan Niehues editor Da Nang, Vietnam conference publication lewis-etal-2015-applying https://aclanthology.org/2015.iwslt-papers.2/ 2015-dec 3-4 126 134