Lexical Micro-adaptation in Statistical Machine Translation Micro-adaptation lexicale en traduction automatique statistique

We introduce a generic framework in Statistical Machine Translation (SMT) in which lexical hypotheses, in the form of a target language model local to the input sentence, are used to guide the search for the best translation, thus performing a lexical microadaptation. An in- stantiation of this framework is presented and evaluated on three language pairs, where these auxiliary hypotheses are derived through triangulation via an auxiliairy language. Our first ex- periments consider nine auxiliary languages, allowing us to measure their individual contribu- tion. We then combine all their hypotheses through a decoding by consensus. Our experiments show that SMT systems can be improved by automatically produced auxiliary hypotheses.