A Stochastic Approach To Character Growth In Automated Narrative Generation

Abstract

This paper describes the AESOP program, a modular storytelling framework that composes storyboards for stories that emphasize character development. Through a combination of modules designed to represent story elements, such as Characters, Actions, and Conditions. AESOP combines the novelty of stochastic events with the grounding of logic progression. These modules form a unique approach that synthesizes previous storytelling systems. A prototype system was built and tested in cooperation with the Fraunhofer Inst, for Digital Media Technology. Our tests experimented with the inclusion and removal of certain modules from the story origination process, and our results showed that the introduction of our modules improved the perceived quality of the generated stories over random story origination. Though the system has a number of shortcomings in terms of its details, readability, and flexibility, it makes great strides in automated story origination.

Publication Date

1-1-2017

Publication Title

FLAIRS 2017 - Proceedings of the 30th International Florida Artificial Intelligence Research Society Conference

Number of Pages

152-157

Document Type

Article; Proceedings Paper

Personal Identifier

scopus

Socpus ID

85029538465 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/85029538465

This document is currently not available here.

Share

COinS