Hang Xu

A new method enhances map creation for self-driving cars.

A new method enhances safety features in multimodal AI systems without extensive training.

This model improves object detection by generating detailed labels for various objects.

JointDreamer enhances 3D generation from text with improved view consistency and quality.

EasyControl simplifies video creation with flexible input methods and enhanced quality.

A new framework enhancing the understanding of images and text together.

EMOVA enhances human-computer interaction through emotional expression.

VidMan teaches robots to learn tasks by watching videos, improving their skills effectively.

AtomThink helps computers solve math problems through step-by-step reasoning.

Robots learn to grasp objects quickly with SparseGrasp, using minimal images.

Agents seek cooperation in changing environments to find optimal solutions.