Quest: a Model of Question Answering

We have developed a computer model of human question answering, called QUEST. QUEST simulates tile answers that people produce when they answer different types of questions, such as why, how, when, where, what-if, and yes/no verification questions. When QUEST answers a particular question, the model identifies relevant information sources and taps information within each source. Each information source is a package of world knowledge that is organized in the form of a "conceptual graph structure" containing nodes and relational arcs. The question answering (Q/A) procedures operate on these structures systematically, pursuing some paths of ares, but not others, depending on the question category. The success of QUEST in simulating human question answering depends critically on an appropriate organization of world knowledge structures as well as an appropriate specification of the Q /A procedures that operate on the structures. The computat ional foundations of QUEST were inspired by models of question answering in artificial intelligence and computat ional linguistics [1-8]. In these models, text and world knowledge are organized as s tructured databases, such as semantic networks [9-11], conceptual dependency theory graphs [12], or conceptual graphs [13]. The Q / A procedures access these information sources and search through the structures systematically by traversing particular categories of arcs. Such models of question answering in AI or computat ional linguistics are regarded as computat ional ly sufficient if they can generate all nodes from the information sources that are relevant to particular questions. QUEST is similar to these models in that it aspires to be a computat ional ly sufficient model of question answering. QUEST was intended to be a psychological model of question answering. It was therefore designed to be psychologically plausible in addition to being computationally sufficient. Tha t is, it was developed under the additional constraint that the answers that QUEST produces should be the same as the answers that adults typically produce. Previous studies have reported the extent This research was funded by grants awarded to the first author by the Office of Naval Research (N0001,I-88-K-0110 and N00014-90-J-1492) and a grant awarded to the second author by the Air Force Office of Scientific Research (88-0063). Requests for reprints should be sent to Arthur C. Graesser, Department of Psychology, Memphis State University, Memphis, TN 38152 U.S.A. Typeset by .A~Vt~5-TEX