Multimodal Emotion Prediction Using Reinforcement Learning

Our aim is to build an ensemble model that is based on reinforcement learning, and capable of capturing the user specific weight of each modality (visual,audio,text) in the emotional status.