Jiaming Liu
Papers
2
Total Citations
5
H-Index
2
About
Jiaming Liu is an emerging researcher working at the intersection of 3D scene understanding, robot perception, and embodied AI. His work addresses fundamental challenges in enabling machines to perceive and interact with three-dimensional environments, with a particular focus on bridging the gap between 2D foundation models and 3D spatial reasoning. Among his notable contributions, Liu's 2025 paper "Lift3D Policy" introduces a framework for adapting powerful 2D foundation models to support robust 3D robotic manipulation — a critical step toward robots that can reason about complex spatial relationships and execute precise physical interactions. Complementing this, his work on "RenderOcc" tackles the computationally and financially demanding problem of 3D occupancy prediction for autonomous driving and robot perception, proposing an innovative approach that leverages 2D rendering supervision to reduce reliance on costly 3D voxel-space annotations. While still early in his career, with his cited works accumulating initial traction in the research community, Liu's contributions reflect a thoughtful and technically sophisticated approach to core problems in embodied intelligence and scene understanding — areas of rapidly growing importance as autonomous systems become increasingly prevalent in real-world applications.
Research Focus
Key Achievements
Top Papers
- 1
- 2