Abstract

As artificial intelligence systems approach human-level performance across a wide variety of domains, creative domains remain a critical yet underexplored area of research. Unlike traditional AI tasks with well-defined objectives, creative tasks rely on inherently subjective measures, making them particularly challenging for machines that lack human experiential context. This dissertation establishes the fundamental importance of creativity in AI systems and presents methods for improving and evaluating the creative capabilities of LLMs on the pathway toward AGI. This work addresses and examines key limitations in current LLM creativity through model innovations, development frameworks, and comparative studies. The dissertation develops novel architectural approaches including an external knowledge integration mechanism inspired by human symbolic memory and black-box controllable text generation methods. Additionally, it presents a game-based development framework for creative agents and an automated knowledge extraction pipeline for story generation systems. Through comparative studies across multiple creative domains, this research provides empirical insights into the capabilities and limitations of LLMs in creative tasks. The work establishes evaluation frameworks for comparing general-purpose LLMs against traditional computational creativity systems and develops methods for assessing the effectiveness of various prompting techniques in creative text generation. These investigations show that LLM-generated creative artifacts are generally preferred over those from traditional CC systems, advanced prompting techniques (e.g., OPRO, chain-of-thought) do not significantly outperform basic prompting, and that LLM-based automatic evaluation is limited. Finally, this work reflects on the advantages and limitations of enhancing LLM creativity and provides directions for future work.

Degree

PhD

College and Department

Computer Science; Computational, Mathematical, and Physical Sciences

Rights

https://lib.byu.edu/about/copyright/