Understanding Goal-Based Stories.
Abstract
Reading requires reasoning. A reader often needs to infer connections between the sentences of a text and must therefore be capable of reasoning about the situations to which the text refers. People can reason about situations because they posses a vast store of knowledge which they can use to infer implicit parts of a situation from those aspects of the situation explicitly described by a text. PAM (Plan Applier Mechanism) is a computer program that understands stories by reasoning about the situations they reference. PAM reads stories in English and produces representations for the stories that include the inferences needed to connect each story's events. To demonstrate that it has understood a story, PAM answers questions about the story and expresses the story from several points of view. PAM reasons about the motives of a story's characters. Many inferences needed for story understanding are concerned with finding explanations for events in the story. PAM has a great deal of knowledge about people's goals which it applies to find explanations for the actions taken by a story's characters in terms of that character's goals and plans.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1978
- Accession Number
- ADA062629
Entities
People
- Robert Wilensky
Organizations
- Yale University