A comparison of several approximate algorithms for finding multiple (N-best) sentence hypotheses
ICASSP '91 Proceedings of the Acoustics, Speech, and Signal Processing, 1991. ICASSP-91., 1991 International Conference
Communicative facial displays as a new conversational modality
CHI '93 Proceedings of the INTERACT '93 and CHI '93 Conference on Human Factors in Computing Systems
Hi-index | 0.00 |
In this paper, a continuous speech recognition system, "niNja" (Natural language INterface in JApanese), is presented. Efficient search algorithms are proposed to get high accuracy and to reduce the required computations. First, an LR parsing algorithm with context-dependent phone models is proposed. Second, scores of the same phone models in different hypotheses at the phone-level are represented by the single score of the best hypothesis. The system is tested for the task with 113 word vocabulary, word perplexity 4.1. It produces sentence accuracy of 97.3% for the 10 open speakers's 110 sentences and the error reduction is as much as 77% comparing with the case using context independent phone models.