... considers both image and text while classifying hate speech. All of these works utilized feature fusion methods, but did not simultaneously consider all textual, visual and acoustic features in hate video. Multimodal learning refers to building models that enable machines to learn information from...