Jingbo Wang

Real-world tennis serving by humanoid robots is now possible without motion capture, thanks to a novel adaptive framework that learns directly from video.

By leveraging a structured latent space, RoboStriker achieves superior tactical performance in humanoid boxing, outperforming traditional methods that struggle with physical feasibility.

Query-adaptive reasoning in 3DGS can reduce target-reference confusion by over 30%, reshaping how we approach open-vocabulary segmentation tasks.

Coordinated scaling of Behavior Foundation Models can enhance humanoid robot control performance, achieving up to 82% error reduction in real-world tasks.

ReactiveBFM enables humanoids to achieve zero-shot moving target reaching with unprecedented agility and real-time adaptability.

RoboNaldo slashes free-kick shot errors by nearly 50% while achieving ball speeds that rival professional soccer players.

By explicitly modeling and constraining human behavior within an epidemic model, SL-BiLEM achieves a remarkable 20x reduction in out-of-distribution performance degradation compared to purely neural approaches when predicting the impact of novel policies.

Robots can now learn flexible, geometry-free interactions with objects directly from video, sidestepping the need for laborious 3D modeling or complex retargeting.

Feel what the robot feels: a new glove lets human operators experience high-resolution tactile feedback during dexterous teleoperation, dramatically improving performance in contact-rich tasks.