Roberto presents at Prof. Kristen Grauman’s group on “What can robots learn from videos about long-horizon activities?” comprising the work on Paprika (CVPR23) and MaskViT (ICLR23).