Application of generative artificial intelligence for epidemic-based modelling

• Main Author: Villaplana Hannah Danielle Ladera
• Other Authors: Cai Wentong
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Computer and Information Science; Generative AI
• Online Access: https://hdl.handle.net/10356/174974
• Institution: Nanyang Technological University

Epidemic models have become increasingly essential, especially in the wake of the recent COVID-19 pandemic, emphasising the crucial role of human behaviour in the spread of disease. There has been a recent rise in the usage and popularity of Generative Artificial Intelligence (AI), such as ChatGPT especially with its ability to mimic human behaviour. This final year project explores the novel application of Generative AI, aiming to overcome the challenge of incorporating nuanced human behaviour in epidemic models. By leveraging ChatGPT, this study seeks to enhance the cognitive realism of agents within an Agent-Based Model (ABM). In this approach, each agent’s actions and behaviour are decided by ChatGPT through a behaviour grid. Various experiments were conducted with different contexts revealing that the simulations successfully generated Susceptible-Infected-Recovered (SIR) graphs and social contact matrices that closely mimic observed epidemic patterns. Furthermore, this project showed the potential of using Random Forests in complementing ChatGPT data to identify the most influential factors in its decision-making. However, the project also highlights the double-edged nature of prompt engineering in ChatGPT, highlighting its sensitivity. The findings underscore the need for careful consideration and refinement in utilising generative AI for modelling complex systems. This project not only contributes to the advancement of epidemic modelling but also underscores the versatility of generative AI in enhancing the cognitive realism of agent-based simulations. As the world grapples with the aftermath of a global pandemic, the imperative for robust epidemic models becomes more pronounced than ever, and the integration of advanced technologies like Generative AI becomes pivotal in addressing future health crises.