Abstract

Topical segmentation is a basic tool for information access to audio records of meetings and other types of speech documents which may be fairly long and contain multiple topics. Standard segmentation algorithms are typically based on keywords, pitch contours or pauses. This work demonstrates that speaker initiative and style may be used as segmentation criteria as well. A probabilistic segmentation procedure is presented which allows the integration and modeling of these features in a clean framework with good results.

Citations

1458	An algorithm for suffix stripping - Porter - 1980
920	Attention, intentions, and the structure of discourse - Grosz, Sidner - 1986
624	Rhetorical structure theory: Toward a functional theory of text organization - Mann, Thompson - 1988
275	TextTiling: Segmenting text into multi-paragraph subtopic passages - Hearst - 1997
191	Topic detection and tracking pilot study: Final report - Allan, Carbonell, et al. - 1998
179	Statistical models for text segmentation - Beeferman, Berger, et al. - 1999
149	The reliability of a dialogue structure coding scheme - Carletta, Isard, et al. - 1997
149	On structuring probabilistic dependencies in stochastic language modeling. Computer Speech and Language - Ney, Essen, et al. - 1994
142	A cache-based natural language model for speech recognition - Kuhn, Mori - 1990
127	Substructure discovery using minimum description length and background knowledge - Cook, Holder - 1994
114	Mixed initiative in dialogue: An investigation into discourse segmentation - Walker, Whittaker - 1990
100	Advances in domain independent linear text segmentation - Choi - 2000
98	The Rhetorical Parsing, Summarization, and Generation of Natural Language Texts’, unpublished - Marcu - 1997
91	The TREC spoken document retrieval track: A success story - Garofolo, Auzanne, et al.
69	Filochat: handwritten notes provide access to recorded converstations - Whittaker, Hyland, et al.
67	Text segmentation by topic - Ponte, Croft - 1997
62	Discourse segmentation by human and automated means - Passonneau, Litman - 1997
59	I'll Get That Off the Audio": A Case Study of Salvaging Multimedia Meeting Records - Moran, Palen, et al. - 1997
52	Advances in automatic meeting record creation and access - Waibel, Bett, et al. - 2001
50	Topic segmentation: Algorithms and applications - Reynar - 2000
49	Document and passage retrieval based on hidden markov models - Mittendorf, Schäuble - 1994
42	Document Expansion for Speech Retrieval - Singhal, Pereira - 1999
39	Acoustic indicators of topic segmentation - Hirschberg, Nakatani - 1998
39	A hidden Markov model approach to text segmentation and event tracking - Yamron, Carp, et al. - 1998
33	Linear segmentation and segment significance - Kan, Klavans, et al. - 1998
20	Adaptive vocabularies for transcribing multilingual broadcast news - Geutner, Finke, et al. - 1998
12	New developments in automatic meeting transcription - Yu, Tomokiyo, et al. - 2000
12	DIASUMM: Flexible summarization of spontaneous dialogues in unrestricted domains - Zechner, Waibel - 2000
10	Interactional Dominance in Dyadic Communication – A Presentation of Initiative–Response Analysis - Linell, Gustavsson, et al. - 1988
6	Shallow discourse genre annotation in callhome spanish - Ries, Levin, et al. - 2000
4	neural network based speech act classification - Ries, “HMM - 1999
3	Gaze Cues for Discourse Segmentation - Gesture
3	Activity detection for information access to oral communication - Ries, Waibel - 2001
2	The effects of room acoustics on MFCC speech parameters - Pan, Waibel - 2000
1	barbara corpus of spoken american english part-i - Santa - 2000
1	lexicon, speech and transcripts - spanish - 1996
1	Prosody modeling for automatic sentence and topic segmentation from speech - Shriberg, Stolcke, et al. - 2000