Abstract. We present in this report two main approaches to cross-language information retrieval based on the exploitation of multilingual corpora to derive cross-lingual term-term correspondences. These two approaches are evaluated in the framework of the multilingual-4 (ML4) task.