Multipass algorithm for acquisition of salient acoustic morphemes

We are interested in spoken language understanding within the domain of automated telecommunication services. Our current methodology involves training statistical language models from large annotated corpora for recognition and understanding. Since the transcribing of large speech corpora is a resource consuming task, we are motivated to exploit speech without transcriptions. In particular, we learn the semantic associations for a task exploiting only phone-based sequences from the output of a task-independent ASR-system. In this paper we present a new multipass algorithm for acquiring salient phone sequences from untranscribed speech corpora and evaluate their utility for the HMIHY task. Compared to our previous strategy, this algorithm is shown to produce improved call-classification results wh ile reducing up to 7-fold the number of salient phone-sequences selected for training.