ISCA Archive Eurospeech 1989
ISCA Archive Eurospeech 1989

Sentence-hypotheses generation in a continuous-speech recognition system

Volker Steinbiss

In this paper, the dynamic-programming algorithm for continuous-speech recognition is modified in order to obtain a top-N sentence-hypotheses list instead of the usual one sentence only. The theoretical basis of this extension is a generalization of Bellman's principle of optimality. Due to the computational complexity of the new algorithm, a sub-optimal variant is proposed, and experimental results within the SPICOS system are presented.