The semantic component of the speech understanding system being developed jointly by SRI and SDC rules out phrase combinations that are not meaningful and produces semantic interpretations for combination that are.
Natural language output can be generated from semantic nets by processing templates associated with concepts in the net. A set of verb templates is being derived from a study of the surface syntax of some 3000 English verbs.
This paper describes part of the discourse component of a speech understanding system for task-oriented dialogs, specifically, a mechanism for establishing a focus of attention to aid in identifying the referents of definite noun phrases.
Two important problems in speech understanding are how to effectively integrate multiple sources of knowledge within the system and how to control the activities of the system to arrive at appropriate interpretations for utterances.
This paper describes the current status of research being performed by Stanford Research Institute on the development of a speech understanding system capable of engaging a human operator in a conversation about a specific task domain.
This paper develops the following strategy: to achieve two goals simultaneously, develop a plan to achieve one of them and then modify that plan to achieve the second as well.
This paper presents some programming facilities for modeling the semantics of a task domain and for describing the situations that occur in that domain as a task is being carried out.
Summary of the present state of research in scene analysis. It identifies fundamental information-processing principles relevant to representation and use of knowledge in vision and traces limitations of existing programs to compromises of these principles necessitated by extant processors.
