Title: Neurosymbolic Automated Story Generation

Lara J. Martin

https://laramartin.net

PhD Candidate in Human-Centered Computing

School of Interactive Computing

Georgia Institute of Technology

Date: Monday, December 7, 2020

Time: 5-8pm EST

Location (Bluejeans): https://gatech.bluejeans.com/5976444105

Committee:

Dr. Mark Riedl - Advisor, Georgia Institute of Technology, School of Interactive Computing

Dr. Ayanna Howard - Georgia Institute of Technology, School of Interactive Computing

Dr. Ashok Goel - Georgia Institute of Technology, School of Interactive Computing

Dr. Devi Parikh - Georgia Institute of Technology, School of Interactive Computing

Dr. Alan W Black - Carnegie Mellon University, Language Technologies Institute

Abstract:

Although we are currently riding a technological wave of personal assistants, many of these agents still struggle to communicate appropriately. Humans are natural storytellers, so it would be fitting if artificial intelligence could tell stories as well. Automated story generation is an area of AI research that aims to create agents that tell "good" stories. Previous story generation systems use planning and symbolic representations to create new stories, but these systems require a vast amount of knowledge engineering. The stories created by these systems are coherent, but only a finite set of stories can be generated. In contrast, very large neural language models have recently made the headlines in the natural language processing community. Though impressive on the surface, these models begin to lose coherence over time. My research looks at neural and symbolic techniques of automated story generation, focusing on the perceived creativity of the generated stories. Here, I define a creative product as one that is both novel and useful. In this dissertation, I created automated story generation systems that improved the novelty and utility (coherence) by separating out semantic event generation from syntactic sentence generation, manipulating neural event generation to become goal-driven, improving syntactic sentence generation to be more novel and useful, and creating a neurosymbolic system that pulls strengths from both neural-only and symbolic-only systems.