Fetching the paper…

Graph2Vid: Flow graph to Video Grounding for Weakly-supervised Multi-Step Localization · Around