University of Twente Student Theses

Login

Integrating a Human Feedback Loop in PCG for Level Design Using LLMs

Koppen, Lucas van (2024) Integrating a Human Feedback Loop in PCG for Level Design Using LLMs.

[img] PDF
4MB
Abstract:With the recent developments of Large Language Models (LLMs) and the technical capabilities of Procedural Content Generation (PCG), new opportunities arise when intersecting these 2 fields. Currently, there has been little to no research on the application of LLMs into PCG. This thesis explores some of these opportunities by researching how the integration of a Human Feedback Loop using LLMs improves PCG techniques in level design for games. To answer this question, a Super Mario Bros generator that uses constructive PCG was taken and the concept has been applied, allowing users to give feedback in the form of text and letting the prototype generate more personalized levels. User evaluations have been conducted to test the performance of the concept and find its strengths and weaknesses. In general, the concept works and shows promise as a tool for level design that can improve efficiency and usability, but using it as a game mechanic for existing games can prove to be challenging. As the capabilities of LLMs improve, the applications for the concept become more interesting within the field of game design. The results of this research show that the application of LLMs in PCG is promising and there are more opportunities for leveraging the interpretive capabilities of LLMs in the game industry.
Item Type:Essay (Bachelor)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Creative Technology BSc (50447)
Link to this item:https://purl.utwente.nl/essays/102052
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page