Nuo Chen

Ph.D. Candidate @ Texas A&M ECE |

[Nuo Chen @Nuo_Chen_6]

🎥 Endoscopy gets messy. SLAM shouldn’t.

🚀 EndoMD-SLAM maintains SLAM stability under optical degradation using memory-driven relocalization and static-transient decomposition.

Paper: https://t.co/hb1T6k9tJX
Project: https://t.co/CMdx6Pe7XQ
Code: https://t.co/DGgtwxCOxT

[Nuo Chen @Nuo_Chen_6]

Excited to share our work, OpenLongTail! 🚀

Our goal is to unlock diverse long-tail driving videos for scalable policy learning by transforming monocular videos into synchronized, pose-grounded surround views.

Paper: https://t.co/SKmvSHTaiJ
Project page: https://t.co/10tGQFz3RQ

[lulinliu @lulinliu192922]

🚗Robust driving policies are bottlenecked by scarce, unusable long-tail data.

🚀OpenLongTail transforms monocular and heterogeneous driving videos into synchronized pose-grounded surround-view assets for scalable VLA training.
- Paper: https://t.co/YnAKvPActR
- Project page: https://t.co/7cXCWbduk1
- Code: https://t.co/rSiAdaFK60

[Nuo Chen @Nuo_Chen_6]

Excited to share our latest work, CrashTwin 🚗💥

A benchmark for quantifying physical violations in world models.

World model rollouts → metric-scale kinematics → physics-grounded metrics.

Paper: https://t.co/3jOEFt7vZS
Project: https://t.co/NM0mRm9BFq

[Zhiwen(Aaron) Fan @zhiwen_fan_]

World models should not just look real. They should obey physics.

CrashTwin stress-tests generative world models on safety-critical crashes, measuring temporal consistency, momentum and energy conservation, and identity stability.

Data + code: https://t.co/3ci6HOAJRK