Interest in flexible and adaptable Machine Translation (MT) systems, based on the language-independent hybrid methods, whose principles ensure easy portability to new language pairs has led to organizing an MT Workshop in the framework of the TSD Conference 2012. These methods attempt to overcome well-known problems of some MT approaches, e.g. bilingual corpora compilation or the laborious creation of the new rules per language pair. Issues of effectively managing multilingual content in a language-independent machine-learning-based methodology are of special importance. This volume contains the proceedings of the Hybrid Machine Translation Workshop organized as a satellite Workshop at the 15 th TSD Conference, held in Brno, Czech Republic on September 3rd, 2012. In the review, eight papers were accepted out of 13 submitted, and appear in this volume.