Submitted by bingxuan li 32 VIS-Ground: Video Interactive Storytelling with Contextual Grounding Google 17 2