A mixed initiative human-computer collaboration in generating children’s stories

Date of Publication

1-7-2019

Document Type

Dissertation

Degree Name

Doctor of Philosophy in Computer Science

Subject Categories

Artificial Intelligence and Robotics

College

College of Computer Studies

Department/Unit

Software Technology

Thesis Adviser

Merlin Teodosia Suarez

Defense Panel Chair

Nelson Marcos

Defense Panel Member

Raymund C. Sison
Susa\n Pancho Festin
Shirley O. Lua
Merlin Teodosia Suarez


Abstract/Summary

Augmented intelligence, combined with computational storytelling, involves the collaboration between a human agent and a software agent to complement and to enhance each other’s abilities in the generation of stories. The extent of control in story planning and generation that is delegated to each of the collaborating entities may vary from one computing environment to another. In this dissertation, end-to-end techniques in story planning and generation were explored to demonstrate three (3) approaches to collaborative storytelling, namely automated, constrained and shared story generation. Three (3) categories of storytelling knowledge needed for the software agent to be an active participant in the collaborative task were also identified. These are the domain knowledge or commonsense ontology, operational knowledge, and linguistic knowledge. Micro and macro evaluation methods were performed to analyze the effect of the variances in planning and generation algorithms to the resulting stories. Micro evaluation looked at each collaboration approach to assess how planning strategies led to the generation of stories that satisfy three dimensions: content, coherence and fluency. A fourth dimension, interaction, was also used in constrained and shared story generation, to assess the extent of collaboration afforded to the participating entities. Macro evaluation provided an overall view across the varying forms of collaboration to assess the variances and consistency in plot structure of the resulting stories. The different storytelling knowledge resources were also evaluated based on the sufficiency and relevance of their content.

Abstract Format

html

Language

English

Format

Electronic

Accession Number

CDTG008258

Keywords

Artificial intelligence; Human-computer interaction; Intelligent agents (Computer software)

Upload Full Text

wf_no

Embargo Period

3-31-2025

This document is currently not available here.

Share

COinS