Speech summarization is a current research area for a broad range of applications, and using many different metholodogies. This special session will both provide a general introduction to the area and bring together researchers working within it for information exchange and discussion.
We invite papers that report current research in speech summarization for any speech genre. This includes (but is not limited to) summarizing meetings, lectures, broadcast news, and spoken dialogue, using output that is spoken, textual, or multimodal. There are currently a wide range of methods that are used for summarization, such as sentence (or dialogue act) compression, sentence extraction, and text rewriting; this session will be eclectic in approach, and so work using any method is acceptable here, including novel ones. Comparisons of speech-specific systems to systems imported from text summarization are especially welcome. Papers could report results from summarization systems, but might also address issues of importance to summarization researchers, such as corpus collection and evaluation, or discuss the differences among different tasks and approaches.