Generating game content with generative language models

• Thesis title in Czech: Generování herního obsahu za použití generativních jazykových modelů
• Thesis title in English: Generating game content with generative language models
• Key words: procedurální generování obsahu|generativní jazykový model|počítačové hry
• English key words: procedural content generation|generative language model|computer games
• Academic year of topic announcement: 2022/2023
• Thesis type: diploma thesis
• Thesis language: angličtina
• Department: Department of Software and Computer Science Education (32-KSVI)
• Supervisor: Mgr. Vojtěch Černý
• Author: hidden - assigned and confirmed by the Study Dept.
• Date of registration: 28.03.2023
• Date of assignment: 11.04.2023
• Confirmed by Study dept. on: 24.05.2023

Guidelines

This thesis aims to create an algorithm that will use generative language models, such as GPT-3/GPT-4, to automatically generate game content. The output of the thesis will be a playable RPG game utilizing this algorithm. Among the content created by the generative language model will be the story of the game, game characters and their phrases. The algorithm will also be used to select sprites from a prepared catalogue.

References

Shaker N., Togelius J., Nelson M.J. 2016. Procedural Content Generation in Games.
Short Tanya X. and Tarn Adams. 2017. Procedural Generation in Game Design.
S. Värtinen, P. Hämäläinen and C. Guckelsberger, "Generating Role-Playing Game Quests With GPT Language Models," in IEEE Transactions on Games, doi: 10.1109/TG.2022.3228480.
White J., Fu Q., Hays S., Sandborn M., Olea C., Gilbert H., et al. A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT. arXiv preprint arXiv:230211382. 2023.