MUTAN: Multimodal Tucker Fusion for Visual Question Answering | IEEE Conference Publication | IEEE Xplore