Hao Luo

TTP enables robots to learn dexterous manipulation from human tactile experiences, achieving unprecedented performance in complex tasks.

UniGP reveals that joint training of controllable generation and dense prediction can significantly enhance performance without the need for complex designs, outperforming specialized models.

EDA not only corrects the current memory write but also actively removes outdated information, leading to superior performance in long-context scenarios.

Seemingly impressive VLA performance on robotic benchmarks crumbles when stress-tested with causal interventions, exposing a reliance on brittle shortcuts rather than genuine embodied reasoning.

Stop guessing how long LLM outputs will be – modeling the *distribution* of possible lengths slashes latency by 2x and boosts throughput by 40%.

Diffusion models can now generate high-fidelity virtual contrast-enhanced CT scans from non-contrast scans, potentially reducing the need for invasive contrast agents and radiation exposure.

Forget synthetic data and limited teleoperation: Being-H0 leverages the dexterity and scalability of human hand videos for VLA pretraining, unlocking superior performance in complex manipulation tasks.