Learning Disentangled Multimodal Representations for the Fashion Domain | IEEE Conference Publication | IEEE Xplore