| Abstract | As experiences continue to advance, recognizing, understanding, and responding to affective states is anticipated to become more important. Emotion-aware systems have significant implications for human-computer interaction. By responding to emotional feedback, experiences can become more personalized and improve usability. Virtual reality (VR) allows users to interact with a fictional world capable of evoking various emotions. Combining VR and affect adaptation is likely to assist user outcomes, for example, for learning or psychological benefits. There is a current lack of real-time solutions developed for affect adaptation, which refers to the system controlling the experience based on emotional feedback inference. This paper suggests a framework that uses deep Q-learning with offline training to identify and respond to emotion-related data, and it outlines future research directions. The aim is to summarize and discuss the challenges in affective computing to establish how affect adaptation can be applied to improve applications in VR. |
|---|