Voice to Victory: Modeling In-Game Voice Communication for Team Coordination in MOBA Games

초록

In Multiplayer Online Battle Arena (MOBA) games, effective communication is crucial for coordinating team strategies to achieve victory. The influence of communication on team coordination have been examined by prior studies; however, were limited to non-verbal cues. We propose an end-to-end speech act recognition pipeline that automatically infers speakers’ intent from voice audio, enabling analysis at scale. Additionally, we applied social network analysis to model verbal communication during team coordination in MOBA. By conducting a data collection study on the gameplay of five players in League of Legends (LoL), we examined team communication was predominantly driven by specific players in collaboratively intensive situations, especially during question-answering interaction.

키워드