Joint Visual-Temporal Embedding for Unsupervised Learning of Actions in Untrimmed Sequences | IEEE Conference Publication | IEEE Xplore