Revisiting Disentanglement and Fusion on Modality and Context in Conversational Multimodal Emotion Recognition. (arXiv:2308.04502v1 [cs.CL])

It has been a hot research topic to enable machines to understand human
emotions in multimodal contexts under dialogue scenarios, which is tasked with
multimodal emotion analysis in conversation (MM-ERC). MM-ERC has received
consistent attention in recent years, where a diverse range of methods has been
proposed for securing better task performance. Most existing works treat MM-ERC
as a standard multimodal classification problem and perform multimodal feature
disentanglement and fusion for maximizing feature utility. Yet after revisiting
the characteristic of …

analysis arxiv attention consistent context conversation conversational dialogue diverse emotion emotions fusion hot human human emotions machines multimodal recognition research