Procedural Persona

Procedural personas are computational models designed to simulate human behavior, particularly within interactive systems like video games. Current research focuses on generating these personas using techniques like reinforcement learning (e.g., Go-Explore) and leveraging large language models to incorporate personality traits and behavioral patterns derived from qualitative data analysis. This work aims to improve the realism and diversity of synthetic data for applications such as computer vision and AI-driven game testing, ultimately leading to more effective and personalized user experiences. The ability to accurately predict and generate diverse player personas holds significant implications for game design, AI development, and the creation of more realistic and engaging virtual environments.