Removing human players from the loop: AI-assisted assessment of Gaming QoE

Quality of Experience (QoE) assessment for video games is known for being a heavy-weight process, typically requiring the active involvement of several human players and bringing limited transferability across games. Clearly, to some extent, QoE is correlated with the achieved in-game score, as player frustration will arise whenever realized performance is far from what is expected due to conditions beyond player control such as network congestion in the increasingly prevalent case of networked games. To disrupt the status quo, we propose to remove human players from the loop and instead exploit Deep Reinforcement Learning (DRL) agents to play games under varying network conditions. We apply our framework to a set of Atari games with different types of interaction, showing that the score degradation observed with DRL agents can be exploited in networking devices (e.g., by prioritizing scheduling decisions), reinforcing fairness across games, and thus enhancing the overall quality of gaming experience.