Document Type
Conference Proceeding
Publication Date
2022
Publication Title
Proceedings of The Tenth Annual Conference on Advances in Cognitive Systems (ACS-22)
Abstract
Scripts, knowledge structures defining sequences of events in stereotypical social situations, were traditionally used to simulate the ways in which people can infer unstated details in understanding a story. In this paper, we describe the MUltiple SCRipt AcTivator (MUSCRAT), and the Script Combination Applier Mechanism (SCAM), significant enhancements of Cullingford’s Script Applier Mechanism which accomplish two novel aims. One system, MUSCRAT, is able to activate more than one script and use them during a story understanding process. The second system, SCAM, uses scripts for story generation, using script variables as “terminals” for combining two or more scripts together to create complex narrative situation structures. The combinational abilities of these systems are enhanced by the fact that the scripts are represented not using linguistic elements but using decompositions into abstract conceptual primitives. We also discuss how some of the perceived weaknesses of scripts stemming from prior work may be overcome in systems that represent general thinking and reasoning processes as combined instantiations of standardized and generalized memory episodes.
Recommended Citation
McKenzie, Megan; Kilayko, Alexis; Macbeth, Jamie C.; Carter, Scott; Sieck, Katharine; and Klenk, Matthew, "Script Combination for Enhanced Story Understanding and Story Generation Systems" (2022). Computer Science: Faculty Publications, Smith College, Northampton, MA.
https://scholarworks.smith.edu/csc_facpubs/395
Comments
Peer reviewed accepted manuscript.
Proceedings of The Tenth Annual Conference on Advances in Cognitive Systems (ACS-22), Arlington, VA, November 19-22, 2022.