Meeting Corpus Released

^{January 15, 2004}
The ICSI Meeting Corpus has now been released by the Linguistic Data Consortium. This corpus, which consists of 75 natural meetings recorded at ICSI from 2000 to 2002, was created with the intention of providing spontaneous multi-party speech data for use in development of speech recognition technology. Speech researchers at ICSI and many other sites (including our colleagues at IDIAP who are working on related material) are interested in developing technology capable of accurately transcribing multi-party meetings, and defiving higher-level information such as summaries from the meetings. The release of the Meeting Corpus marks the completion of the first stage of this kind of research.

For more information on the Meeting Corpus you can visit the LDC website.