The importance of intention in conversation has been considered by many researchers in artificial intelligence and psychology. However, most models of conversation have been limited to pursuing the transfer of knowledge between the system and a user. We propose that conversational goals can address...

Ausführliche Beschreibung

Bibliographische Detailangaben

Hauptverfasser:	Johnson, Peter N. (VerfasserIn), Robertson, Scott P. (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	[New Haven, Conn.] 1981
Schriftenreihe:	Yale University <New Haven, Conn.> / Department of Computer Science: Research report 206
Schlagworte:	MAGPIE(Multiple Active Goal Processor in Interactive Exchanges) Artificial intelligence Computer communications Computer logic Computerized simulation Data bases Emotions Goal programming Humans Interrogation Man computer interface Natural language Non-radio Communications Psychology Künstliche Intelligenz
Zusammenfassung:	The importance of intention in conversation has been considered by many researchers in artificial intelligence and psychology. However, most models of conversation have been limited to pursuing the transfer of knowledge between the system and a user. We propose that conversational goals can address communication at a number of other levels such as the conversats' emotions, their relationship, and their attitudes. MAGPIE (Multiple Active Goal Processor in Interactive Exchanges) is a computer model of a conversant that acquires and pursues conversational goals at a number of levels, including the goal of seeking dominance in its relationship with the other conversant. At the heart of the program is a set of tracking procedures, each of which monitors a specific level of communication flow in a conversation. These procedures are coupled with a conversational goal planner which generates responses that simultaneously pursue a number of goals. Currently, MAGPIE is able to model a wife during a short marital dispute with her husband. Normative data form human subjects is presented which supports the conversational goals proposed in our analysis. (Author).