Modality to Modality Translation: An Adversarial Representation Learning and Graph Fusion Network for Multimodal Fusion

Sijie Mai, Haifeng Hu, Songlong Xing

Learning joint embedding space for various modalities is of vital importance for multimodal fusion. Mainstream modality fusion approaches fail to achieve this goal, leaving a modal…