What When and Where? Self-Supervised Spatio Temporal Grounding in Untrimmed Multi-Action Videos from Narrated InstructionsBrian ChenNina Shvetsovaet al.2024CVPR 2024Conference paper