Advancing DRL Agents in Commercial Fighting Games: Training, Integration, and Agent-Human Alignment
This work addresses the problem of enhancing player training and engagement in commercial fighting games, representing an incremental advancement in DRL agent integration.
The paper tackled the challenge of prolonged player interaction in fighting games by developing Shūkai, a DRL agent system deployed in Naruto Mobile, which achieved consistent competence across all characters with training on only 13% of them and a 22% improvement in sample efficiency.
Deep Reinforcement Learning (DRL) agents have demonstrated impressive success in a wide range of game genres. However, existing research primarily focuses on optimizing DRL competence rather than addressing the challenge of prolonged player interaction. In this paper, we propose a practical DRL agent system for fighting games named Shūkai, which has been successfully deployed to Naruto Mobile, a popular fighting game with over 100 million registered users. Shūkai quantifies the state to enhance generalizability, introducing Heterogeneous League Training (HELT) to achieve balanced competence, generalizability, and training efficiency. Furthermore, Shūkai implements specific rewards to align the agent's behavior with human expectations. Shūkai's ability to generalize is demonstrated by its consistent competence across all characters, even though it was trained on only 13% of them. Additionally, HELT exhibits a remarkable 22% improvement in sample efficiency. Shūkai serves as a valuable training partner for players in Naruto Mobile, enabling them to enhance their abilities and skills.