VideoBERT: A Joint Model for Video and Language Representation Learning | IEEE Conference Publication | IEEE Xplore