Multimodal Comparable Corpora as Resources for Extracting Parallel Data: Parallel Phrases Extraction

Discovering parallel data in comparable corpora is a promising approach for overcoming the lack of parallel texts in statistical machine translation and other NLP applications. In this paper we propose an alternative to comparable corpora of texts as resources for extracting parallel data: a multimodal comparable corpus of audio and texts. We present a novel method to detect parallel phrases from such corpora based on splitting comparable sentences into fragments, called phrases. The audio is transcribed by an automatic speech recognition system, split into fragments and translated with a baseline statistical machine translation system. We then use information retrieval in a large text corpus in the target language, split also into fragments, and extract parallel phrases. We compared our method with parallel sentences extraction techniques. We evaluate the quality of the extracted data on an English to French translation task and show significant improvements over a state-ofthe-art baseline.