This paper describes the latest Speech-to-Text system developed for the Global Autonomous Language Exploitation ("GALE") domain by Carnegie Mellon University (CMU). This systems uses discriminative training, bottle-neck features and other techniques that were not used in previous versions of our system, and is trained on 1150 hours of data from a variety of Arabic speech sources. In this paper, we show how different lexica, pre-processing, and system combination techniques can be used to improve the final output, and provide analysis of the improvements achieved by the individual techniques.
Bibliographic reference. Metze, Florian / Hsiao, Roger / Jin, Qin / Nallasamy, Udhyakumar / Schultz, Tanja (2010): "The 2010 CMU GALE speech-to-text system", In INTERSPEECH-2010, 1501-1504.