Skip to Main Content
Stanford Research Institute (SRI) is participating in a major program of research on the analysis of continuous speech by computer. The goal is the development of a speech understanding system capable of engaging a human operator in a natural conversation about a specific problem domain. The approach being taken is distinctive in the extent to which it depends on syntactic and semantic processing to guide the acoustic analysis. This correspondence provides a description of the first version of the system, emphasizing the kinds of information that need to be added for effective results.