Search papers, labs, and topics across Lattice.
3
0
5
2
Failure rollouts can transform bad actions into valuable learning opportunities, leading to more robust action generation in World-Action Models.
EgoSteer achieves 9x higher throughput and better accuracy in training dexterous robots using egocentric videos, enabling them to execute complex tasks with remarkable success rates.
Forget brittle visual-history buffers: LoHo-Manip uses a VLM task manager with visual trace prompts to achieve robust long-horizon robotic manipulation through implicit closed-loop replanning.
