Ye Wang

Unlock high-fidelity multimodal image generation and editing with a surprisingly simple framework that leverages pre-trained VLMs and diffusion models, bypassing complex fusion architectures.

Musculoskeletal robots can achieve significantly better locomotion by co-designing morphology and control, simultaneously optimizing muscle strength, velocity, and stiffness, rather than relying on fixed physiological parameters.

Achieve 42% faster LLM inference on satellite networks by cleverly splitting models and compressing activations for efficient communication.

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.