Fast Speaker Adaptation in Automatic Online Subtitling

This paper deals with speaker adaptation techniques well suited for the task of online subtitling. Two methods are briefly discussed, namely MAP adaptation and fMLLR. The main emphasis is laid on the description of improvements involved in the process of adaptation subject to the time requirements. Since the adaptation data are gathered continuously, simple modifications of the accumulated statistics have to be carried out in order to make the adaptation more accurate. Another proposed improvement efficiently employs the combination of fMLLR and MAP. In the case of online adaptation no prior transcriptions of the data are available. They are handled by a recognition system, thus it is suitable to assign a well-applied confidence measure to each of the transcriptions. We have performed experiments focused on the trade-off between the adaptation speed and the amount of adaptation data. We were able to gain a relative reduction of WER 16.2 %.