1 paper
Mengyi Shan, Zecheng He, Haoyu Ma +4
Can a video generation model be repurposed as an interactive world simulator? We explore the affordance perception potential of text-to-video models by teaching them to predict hum…