Siamese Learning with Joint Alignment and Regression for Weakly-Supervised Video Paragraph Grounding | Read Paper on Bytez