Can Cui

Trajectory scoring can be dramatically improved by leveraging a dataset that better captures decision boundaries, leading to enhanced performance in autonomous driving systems.

CoRE reveals that coarse video predictions can effectively guide the identification of fine-grained risk evidence, eliminating the need for detailed annotations.

GlanceWAM achieves 72.2% success on RoboCasa while executing actions 24 times faster than traditional synchronous models, reshaping the landscape of real-time robot learning.

Direct predictions in driving world models can lead to substantial inaccuracies in counterfactual scenarios, revealing a critical gap in current methodologies.

Post-training techniques could be the key to overcoming the limitations of traditional imitation learning in autonomous driving, ensuring safer and more reliable vehicle behavior in complex environments.

Even slight variations in natural language instructions can cause language-driven autonomous driving models to fail dramatically, revealing a critical reliability gap for real-world deployment.

Ditch slow, verbose language-based action planning: this diffusion model directly predicts discrete, kinematically feasible waypoints for faster, more precise autonomous driving.

Finally, a real-world testing platform that can rigorously evaluate full-stack VLM-integrated autonomous driving systems, offering configurable scenarios and closed-loop control.