VL-BERT: Pre-training of Generic Visual-Linguistic Representations - Citation Graph | Papersgraph