Integrating a non-probabilistic grammar into large vocabulary continuous speech recognition

We propose a method of incorporating a non-probabi-listic grammar into large vocabulary continuous speech recognition (LVCSR). Our basic assumption is that the ut-terances to be recognized are grammatical to a sufficient degree, which enables us to decrease the word error rate by favouring grammatical phrases. We use a parser and a hand-crafted grammar to identify grammatical phrases in word lattices produced by a speech recognizer. This information is then used to rescore the word lattice. We measured the benefit of our method by extending an LVCSR baseline sys-tem (based on hidden Markov models and a 4-gram lan-guage model) with our rescoring component. We achieved a statistically significant reduction in word error rate com-pared to the baseline system. 1