Minimalist explanations in knowledge-based systems

Research in discourse comprehension and human-computer interaction indicates that good explanations are usually brief. A system that provides brief explanations, however, must plan for the case where brevity comes at the expense of understanding. Human-to-human dialogue is, to a large part, concerned with conversational repair and question-answer episodes; computer systems need to provide similar fallback techniques to their users. The authors have designed such an explanation system in the context of a knowledge-based critiquing system, LISP-CRITIC. The system provides several levels of explanations, specifically tailored to the user. If the initial, brief explanation is insufficient, the system positions the user at an appropriate point within a more complete, hypertext-based documentation system. Rather than attempting to design a system that can generate a perfect one-shot explanation for any given situation, this approach concentrates on matching the communication abilities provided by current computer technology to the cognitive needs of the human user.<<ETX>>

[1]  B. Chandrasekaran,et al.  Explaining control strategies in problem solving , 1989, IEEE Expert.

[2]  Timothy A. Post,et al.  On the role of prior knowledge and task demands in the processing of text , 1988 .

[3]  Gerhard Fischer,et al.  A Critic for LISP , 1987, IJCAI.

[4]  L. Suchman Plans and situated actions , 1987 .

[5]  Laurence Danlos,et al.  The Linguistic Basis of Text Generation , 1987, EACL.

[6]  W. Kintsch,et al.  Strategies of discourse comprehension , 1983 .

[7]  Kathleen McKeown,et al.  Tailoring Explanations for the User , 1985, IJCAI.

[8]  Gerhard Fischer,et al.  Computer-based critics , 1989, [1989] Proceedings of the Twenty-Second Annual Hawaii International Conference on System Sciences. Volume III: Decision Support and Knowledge Based Systems Track.

[9]  Walter Kintsch,et al.  The Representation of Knowledge and the use of Knowledge in Discourse Comprehension1 1This research was supported by National Institute of Mental Health Grant MH 15872. , 1989 .

[10]  Raymond McCall,et al.  Design environments for constructive and argumentative design , 1989, CHI '89.

[11]  H. Simon,et al.  Perception in chess , 1973 .

[12]  George R. Klare,et al.  The measurement of readability , 1963 .

[13]  Johanna D. Moore Responding to :20HUH?”: answering vaguely articulated follow-up questions , 1989, CHI '89.

[14]  Donald A. Waterman,et al.  An Explanation Facility for the ROSIE (Tradename) Knowledge Engineering Language , 1986 .

[15]  Edmund H. Weiss Breaking the grip of user manuals , 1988, ASTR.

[16]  G. Fischer,et al.  The enhancement of understanding through visual representations , 1986, CHI '86.

[17]  Wilfred J. Hansen,et al.  Reading and writing with computers: a framework for explaining differences in performance , 1988, CACM.

[18]  Gerhard Fischer,et al.  Communication requirements for cooperative problem solving systems , 1990, Inf. Syst..

[19]  William R. Swartout,et al.  XPLAIN: A System for Creating and Explaining Expert Consulting Programs , 1983, Artif. Intell..

[20]  Thomas Mastaglio,et al.  User modelling in computer-based critics , 1990, Twenty-Third Annual Hawaii International Conference on System Sciences.

[21]  Wayne S. Murray,et al.  Dialogue with machines , 1988, Cognition.