Infinite_Loop_v1_Duncan_2024_Using_ReinforcementLearning_to_Train_In_game_Non_Player_Characters

Infinite_Loop_v1_Duncan_2024_Using_ReinforcementLearning_to_Train_In_game_Non_Player_Characters

Abstract: The video game industry is continuously evolving, driven by technological advancements, offering opportunities for more immersive gaming experiences. This research explores the integration of machine learning in game design, focusing on applications for in-game artificial intelligence (AI). Traditional game AI design relies on predefined algorithms and scripted behaviors. This often results in predictable AI and static game play. Here I investigate how machine learning can create AI systems that dynamically adapt to a player’s skill level to enhance engagement and challenge. I first present a literature review to learn about current uses of machine learning in video games. Then I describe a prototype game using the Unity ML framework. I use the creation and success of this prototype to test the feasibility of incorporating AI into video game development.

Leave a Reply