|
I am a third-year Computer Science Ph.D. student at Purdue University, advised by Prof. Raymond A. Yeh. I hold a Master of Science in Computer Science degree from University of Illinois at Urbana-Champaign, where I worked with Mr. Yunze Man and Prof. Liangyan Gui . I received my B.S.E. degree in Data Science at University of Michigan with a minor in Mathematics. I was fortunate to work with Dr. Junming Zhang and Prof. Matthew Johnson-Roberson at the UM Ford Center for Autonomous Vehicles (FCAV). I also hold a B.E. degree in Electrical and Computer Engineering at Shanghai Jiao Tong University (SJTU) (Dual Degree Program). Email / CV / Google Scholar / Github / Twitter |
|
|
|
| Mar, 2025 | I will join Mitsubishi Electric Research Laboratories as a research intern in Summer 2025. |
|---|---|
| Nov, 2024 | I am selected as one of the Top Reviewers of NeurIPS 2024. |
| Sep, 2024 | Our paper D-LISA is accepted to NeurIPS 2024. |
| Aug, 2023 | I start as a Computer Science Ph.D. student at Purdue University. |
| Feb, 2023 | Our paper HyperPC is accepted to CVPR 2023. |
|
I am broadly interested in computer vision, machine learning, and robotics, with a focus on 3D vision and multi-modal learning in open-world settings. My current work focuses on open-vocabulary 3D object recognition. My previous research has explored (i) affordance generalization, (ii) 3D visual grounding and (iii) point cloud completion. My research goal is to develop computer vision techniques for real-world embodied agents and autonomous systems, enhancing their capability to perceive and reason effectively in uncertain environments. |
|
Haomeng Zhang, Chiao-An Yang, Raymond A. Yeh NeurIPS 2024 arXiv / Project Page / Code Our proposed model D-LISA has a novel vision module that allows for a dynamic number of proposal boxes and extracts features from dynamic viewpoints per scene. Furthermore, we propose a fusion module that is spatially aware with explicit language conditioning. |
|
Junming Zhang, Haomeng Zhang, Ram Vasudevan, Matthew Johnson-Roberson CVPR 2023 arXiv / Project Page / Code We propose a hyperspherical module which could be inserted into any existing Encoder-Decoder structure and consistently improve the point cloud completion result in both single-task and multi-task learning. |
|
Conference Reviewer: CVPR, ICCV, NeurIPS, ICLR, ICML.
|
|
|
|