2000 · 23 citations · 16 references
ComputingComputer ScienceComputational LinguisticsInformation RetrievalNatural Language ProcessingAnnotationCorpus AnalysisCorpus AnnotationMultilingual Natural Language ProcessingMultilingual ProcessingTopic ModelingCross-language RetrievalCross-language Topic DetectionCross-language TransferEvaluation Of Cross-language RetrievalTopic LabelingTopic Tracking
Topic Detection and Tracking (TDT) refers to automatic techniques for discovering, threading, and retrieving topically related material in streams of data. Newswire and broadcast news are the canonical sources. In 1999, TDT research was extended from English to Chinese, and carefully annotated multilingual corpora were created. Researchers devised clever approaches to the cross-language challenge, and formal performance evaluations yielded very promising results. This paper outlines the 1999 research tasks, corpora, evaluation procedures, technical approaches, and results. The multilingual, multimedia research and evaluations are continuing in 2000 and 2001 under the DARPA TIDES program.
16
The DET curve in assessment of detection task performance
Alvin F. Martín, George R. Doddington, Terri Kamm et al. · 1997 · 1.4K citations
A study of retrospective and on-line event detection
Yiming Yang, Tom Pierce, Jaime Carbonell · 1998 · 731 citations
Statistical Models for Text Segmentation
Doug Beeferman, Adam Berger, John Lafferty · Machine Learning · 1999 · 687 citations · Full text
Prosody-based automatic segmentation of speech into sentences and topics
Elizabeth Shriberg, Andreas Stolcke, Dilek Hakkani‐Tür et al. · Speech Communication · 2000 · 424 citations · Full text