Summarization of Spoken Language — Challenges, Methods, and Prospects

While the field of summarizing written texts has been explored for many decades, gaining significantly increased attention in the last five to ten years, summarization of spoken language is a comparatively recent research area. As the amount of spoken audio databases is growing rapidly, however, we predict that the need for high quality summarization of information contained in this medium will rise substantially. Summarization of spoken language may also aid the archiving, indexing, and retrieval of various records of oral communication, such as corporate meetings, sales interactions, or customer support. The purpose of this paper is to place summarization of spoken language in the context of general summarization research, describe its main challenges which are added on top of the already challenging area of written text summarization, describe past and current approaches and systems, and finally provide a tentative outlook on future directions in research and development of spoken language summarization systems.