This tutorial has been cancelled
Gareth J. F. Jones (Dublin City University) [Short Bio]
Martha Larson (Delft University of Technology) [Short Bio]
http://www.cdvp.dcu.ie/searchingspeech/
Abstract:
Speech search is concerned with the retrieval of spoken content from collections of speech or multimedia data. The key challenges raised by speech search are indexing via an appropriate process of speech recognition and efficiently accessing specific content elements within spoken data. The specific limitations of speech recognition in terms of vocabulary and word accuracy mean that effective speech search often does not reduce to application of information retrieval to speech recognition transcripts. Although text information retrieval techniques are clearly helpful, speech retrieval involves confronting issues less apt to arise in the text domain, such as high levels of noise in the indexed data and a lack of a clearly defined unit of retrieval. This tutorial will provide researchers in information retrieval with an introduction to the challenges and technologies of spoken content search. It is designed for a broad audience including students undertaking PhD research in spoken content access as more experienced researchers looking to extend or update their knowledge. The tutorial will review the history of spoken content search to date, its component technologies (including a summary introduction to speech recognition), its relationship to text information retrieval, critical system design issues, domains of application, and issues of interaction with spoken content to support efficient access to specific content within spoken content. We will overview initiatives for evaluation of speech search, available resources to support research and development in the area of spoken content retrieval, and the open challenges that we see going forward. [Back to top]
Bios:
Gareth Jones
Gareth Jones is a faculty member at Dublin City University Ireland, where he leads research in several areas of information retrieval, including searching speech content. He has been working on tasks involving spoken content access for more than 15 years. His research in this area began with searching voice mail in the pioneering Video Mail Retrieval using Voice project at the University of Cambridge, and has since included work on a diverse range of applications including broadcast news retrieval, cross-language speech search, and exploitation of speech content for multimedia indexing and search. He is currently leading a new project on improved indexing for speech search funded by Science Foundation Ireland. His contributions to the spoken content retrieval community include coordinating various speech search tracks in the CLEF benchmark campaign since 2002, and participating in the recently launched series of workshop series on Searching Spontaneous Conversation Speech (SSCS). [Back to top]
Martha Larson
Martha Larson is a researcher in the area of speech search and multimedia retrieval at the Delft University of Technology. She is a co-organizer of VideoCLEF, the track of the Cross-Language Evaluation Forum (CLEF) benchmark campaign that is devoted to cross-language access to video collections and is currently involved with organizing the Third Workshop on Searching Spontaneous Continuous Speech (SSCS 2009) to be held in conjunction with ACM MultiMedia. Her scientific interest in speech search technology took root as an intern at IBM Watson Research Center further developed while working as a researcher at the Fraunhofer Institute in Germany. Research projects she has worked on have developed techniques for speech-based access to broadcast news, radio talk shows, parliamentary speeches and podcasts. Currently, she works on video retrieval in the project PetaMedia, EU-Network of Excellence dedicated to improving multimedia access technology by integrating multimedia content analysis and user-based metadata sources. [Back to top]
This tutorial has been cancelled
