Acoustic indicators of topic segmentation.

作者: Christine H. Nakatani , Julia Hirschberg

DOI:

关键词:

摘要: The segmentation of text and speech into topics subtopics is an important step in document interpretation. For text, formatting information, such as headings paragraphing, available to aid this endeavor, although information by no means su cient. speech, the task even more di cult. We present results application machine learning techniques automatic identi cation intonational phrases beginning ending 'topics' determined independently annotators for two corpora | Boston Directions Corpus Broadcast News (HUB-4) DARPA/NIST database.

参考文章(7)
John F. Pitrelli, Julia Hirschberg, Mary E. Beckman, Evaluation of prosodic transcription labeling reliability in the tobi framework. conference of the international speech communication association. ,(1994)
Jacques M. B. Terken, Marc Swerts, Ronald Geluykens, Prosodic correlates of discourse units in spontaneous speech conference of the international speech communication association. pp. 421- 424 ,(1992)
Janet Breckenridge Pierrehumbert, The phonology and phonetics of English intonation ,(1987)
Michael D. Riley, Some applications of tree-based modelling to speech and language Proceedings of the workshop on Speech and Natural Language - HLT '89. pp. 339- 352 ,(1989) , 10.3115/1075434.1075492
Barbara J. Grosz, Christine H. Nakatani, Julia Hirschberg, Discourse Structure in Spoken Language: Studies on Speech Corpora national conference on artificial intelligence. ,(1995)
Julia Hirschberg, Christine H. Nakatani, A prosodic analysis of discourse segments in direction-giving monologues Proceedings of the 34th annual meeting on Association for Computational Linguistics -. pp. 286- 293 ,(1996) , 10.3115/981863.981901