Multi-Modal Temporal Convolutional Network for Anticipating Actions in Egocentric Videos | IEEE Conference Publication | IEEE Xplore