Reusing Multimodal Agents: Efficient Evaluation of Sound Designs in Games
A novel framework for efficiently evaluating in-game sound designs using multimodal learning and reinforcement learning is introduced, which can detect differences in agent performance caused by degraded or altered sound design.