The Janus-III Translation System: Speech-toSpeech Translation in Multiple

The Janus-III system translates spoken languages in limited domains. The current re- search focus is on expanding beyond tasks involving a single limited semantic domain to significantly broader and richer domains. To achieve this goal, The MT components of our system have been engineered to build and manipulate multi-domain parse lattices that are based on modular grammars for multiple semantic domains. This approach yields solutions to several problems including multi- domain disambiguation, segmentation of spoken utterances into sentence units, modularity of system design, and re-use of earlier systems with incompatible output.