I am Ziqi Pang, a learner, researcher, and engineer on robotics and artificial intelligence, with particular experience and interest in robot perception (Computer Vision and SLAM). I am actively seeking to enter a Ph.D. program for 2021 Fall. Currently, I have been honored to receive the offer of a prestigious research program M.S. in Robotics from the School of Computer Science, Carnegie Mellon University.
For the time being, I am working with Dr. Naiyan Wang on autonomous driving at TuSimple, a leading player in truck autonomous driving. Prior to this, I received a B.S. in Computer Science with honor from Peking University. During my undergraduate years, I was fortunate to collaborate with Prof. Martial Hebert at CMU, Prof. Shiliang Zhang at PKU, and Prof. Yuxiong Wang at UIUC.
I am generally interested in all aspects related to building up functional robotics systems. In particular, I took the most passion in answering the following questions:
3D Robot Perception: Perception enables robot applications to perceive the surroundings and make interactions. Therefore, I wish to develop perception modules with strong 3D awareness and ability to combine multiple sensors.
End-to-end (Perception, Planning, Control) Learning: Similar to CNNs free people from feature engineering, end-to-end learning the pipeline of perception, planning, and control poses an attractive vision; thus, it is important in academics. However, engineers in the industry prefer building modules independently, as this makes debugging and analysis more clear. Therefore, I am passionate about end-to-end representation for robots while decoding information for debugging and interpretation.
Learning from Human Experiences: Imitating human behaviors from videos or demonstrations are critical for complex robotics tasks. Therefore, I seek to find proper interfaces and policies to transfer human knowledge onto robots.
Publication and Research Highlights
For details of each project, please refer to the Publication page.
Model-free Vehicle Tracking and State Estimation in Point Cloud Sequences, Ziqi Pang, Zhichao Li, Naiyan Wang. (Currently under the patent application in the company, soon appear on arxiv.)
Unlocking the Full Potential of Small Data with Diverse Supervision, Ziqi Pang*, Zhiyuan Hu*, Pavel Tokmakov, Yuxiong Wang, Martial Hebert (* indicates equal contribution, currently under review)