Attention-Based Multimodal Fusion for Video Description | IEEE Conference Publication | IEEE Xplore