Papers
8
Total Citations
162
H-Index
4
About
Jiazhao Zhang is a robotics and embodied AI researcher whose work spans robotic perception, manipulation, and navigation — areas at the intersection of computer vision, 3D scene understanding, and autonomous systems. He is perhaps best known for **GraspNeRF** (2023, 91 citations), a pioneering framework that leverages generalizable Neural Radiance Fields to enable 6-DoF grasp detection for transparent and specular objects — a notoriously difficult challenge where conventional depth cameras fail. This contribution marked the first multiview RGB-based solution to this problem, earning significant recognition from the robotics community. His earlier work on active scene understanding through online semantic reconstruction (2019, 38 citations) demonstrated his long-standing interest in intelligent robot exploration and semantic 3D mapping. More recently, Zhang has expanded into mobile manipulation with GAMMA (2024, 20 citations), addressing real-time grasping pose estimation during robot approach, and into vision-and-language navigation through the NaVid series, which incorporates video-based reasoning and 4D spatial intelligence. His 2025 platform RoboVerse further reflects a growing commitment to scalable, generalizable robot learning infrastructure. Across these contributions, Zhang has established himself as a versatile and impactful voice in modern embodied AI research.
Research Focus
Key Achievements
Top Papers
- 1
- 2Active Scene Understanding via Online Semantic Reconstruction38 citations · 2019
- 3
- 4
- 5Active Scene Understanding via Online Semantic Reconstruction3 citations · 2019
- 6
- 7
- 8