About

Huan-ang Gao（高焕昂）

Beijing, China

[ Email | GitHub | Google Scholar ]

Short Bio

  • Ph.D. Student @ Dept. of Computer Sci. & Tech. & Institute for AI Industry Research, Tsinghua University (2024.8-Now)

    • Advised by Prof. Ya-Qin Zhang, co-advised by Prof. Hao Zhao.
    • Research interest lying in Generative Simulation for Embodied AI.

  • B. Eng. @ Dept. of Computer Sci. & Tech., Tsinghua University (2020.9-2024.6)

    • GPA 3.98 / 4.00. Ranked #1 out of 204 in the Dept., with 41 courses receiving A or A+ grade.
    • Selected A+ courses: Artificial Neural Networks, Deep Reinforcement Learning, Software Engineering, Data Structures

  • Counselor @ Student Association of Sci. & Tech. of the Dept. (2024.8-Now)

    • I served as General Chair of the Association from 2023.6 to 2024.6.
    • I am now responsible for technical training/competition guidance/scientific research resource integration/internship opportunity integration, etc. for undergraduates in DCST, THU and IIIS, THU. If you get campus recruitment & daily internship opportunities, please feel free to contact me as I can share them with students in these two departments :)

News

  • I had wonderful research internship at Lightwheel.AI and Haomo.AI during the summer of 2024 :)
  • I had the privilege of receiving the highest distinction for undergraduates at Tsinghua and within the Dept. of CST in 2023. I also had the honor of being awarded the SenseTime scholarship in 2023, meeting old and new friends at SenseTime, Shanghai at the winter of 2024.

Research

The primary focus of my research is on generative simulation for embodied AI. In the era of building robots with artificial general intelligence (AGI), data is essential as it empowers the development and evaluation of learning-based algorithms. However, acquiring data and labels is more expensive in the settings of robotics, especially considering the failure cost in corner cases. To address this issue, we resort to generative AIs, which have revolutionized the way we approach simulations. My specific research interest lies in:

  • Generative models (paritularly, diffusion models) and their application in different representations, which helps us to formulate the generation and editing of images, point clouds, videos, neural fields and so on.
  • 3D reconstruction and graphics (particularly, NeRFs, 3DGS and 4DGS), which helps us to initialize a realistic scenario for editing.
  • Closed-loop evaluation and training between generative simulators (as world dynamic models) and policies (such as end-to-end autonomous driving systems).

I have published 10+ papers at top-level conferences including ICCV, ECCV, CVPR, NeurIPS, ICLR, CoRL, ICRA and SIGGRAPH (with 6 papers as first or co-first authors). Please see my Google Scholar profile for details. Really welcome fellow researchers to exchange ideas with me :)

Serving Experience

  • Teaching Assistant
    • (30240163) Software Engineering. Compulsory course in CST, THU. (23Spring, 23Fall, 24Spring, 24Fall, 25Spring)
    • (30240551) Digital Logic Experimentation. Compulsory course in CST, THU. (24Spring, 25Spring)
    • (40240354) Computer Organization and Design. Compulsory course in CST, THU. (23Fall)
  • Former General Chair @ Student Association of Sci. & Tech., CST, THU. (2023.5-2024.6)
    • The homepage of our Association is launched here!
    • See our introduction to CS documents here and our Summer Training Camp of CS Skills for freshmen here.
    • We also launched a resource integration website for our undergraduate students, featuring the latest information on recruitment, daily internship opportunities, and M.S. or Ph.D. hiring offers from partner laboratories and cooperations.

Misc

  • Isn’t it cool to create a realistic world brimming with swords and magic within the realm of virtual reality?