Corpus parallèles massifs et traduction Quelles pépites se cachent dans les séquences-pivots des moteurs de traduction phrase-based ?

The author examines examples of “phrases” which were automatically detected by a reference PBMT (Phrase-Based Machine Translation) engine, by processing the European Parliamen’s massive parallel corpora (Europarl). By doing so, beyond technical considerations, he would like to highlight the importance of such phrases for the systematic study of phraseology, as well as that of translation strategies.