Representations of Prosody in Computational Models for Language Processing

Mari Ostendorf, University of Washington

In this talk, we look at the importance of prosody in the processing of speech for information retrieval and human-computer interaction. Prosody provides information complementary to the words in the signal, and can be used to extract structural information in a rich, automatically generated transcript that better serves language processing applications. In particular, we look at three interrelated types of structure (segmentation, prominence and syntax), methods for automatic detection, the benefit of optimizing rich transcription for the target language processing task, and the impact of this structural information in tasks such as parsing, topic detection, information extraction and translation.