A unification-grammar-directed one-pass search algorithm for parsing spoken language

The author presents an efficient parsing algorithm for integrating the search problems both in speech and language processing in general use for spoken language systems (SLSs). The parsing algorithm is an extension of the finite-state-network (FSN)-directed, one-pass search algorithm to one directed by a context-free grammar or a unification-based grammar with retention of the frame-synchronous procedure. This integrating algorithm can approximately find a global optimal sentence hypothesis, and it avoids the overhead of hierarchical systems based on the lattice parsing approach. The computational complexity of this algorithm is proportional to the length of the input utterance. As the search process in speech recognition can directly account for the predictive information, this framework can be extended to the SLS to deal with dynamically varying constraints in a dialogue situation