On the Use of Word Alignments to Enhance Bitext Compression

The amount of information that is stored in digital form in more than one language is growing very fast as a consequence of the globalization. Furthermore, there are countries and supra-national entities whose legislation enforces the translation (and storage) of all the official texts into all their official languages. Two texts that are mutual translations are usually referred to as a bilingual parallel corpus or, in short, as a bitext. Compressing independently the two texts of a bitext is far form efficient, since the information conveyed by both texts, the meaning, is similar. We take advantage of this fact to devise a bitext compression algorithm that compresses and stores the two texts that form a bitext simultaneously. In our approach, a single model is used to represent both bitext components. For this purpose, we define a biword as a pair made of two words, each one from a different text, that are mutual translations in the bitext. This new concept allows one to represent with a single symbol two words with high mutual information.