One of the important observations done during the CLEF 2009 campaign (Ferro and Peters, 2009) related to CLIR was that the usage of Statistical Machine Translation (SMT) systems (eg. Google Translate) for query translation led to important improvements in the cross-lingual retrieval performance (the best CLIR performance increased from ˜55% of the monolingual baseline in 2008 to more than 90% in 2009 for French and German target languages). However, generalpurpose SMT systems are not necessarily adapted for query translation. That is because SMT systems trained on a corpus of standard parallel phrases take into account the phrase structure implicitly
