Gao Jiawei「髙嘉偉」
Hi! I am an undergraduate student in Department of Automation, Tsinghua University.
I am also a member of the "Tong Class",
an AGI program founded by Prof. Song-Chun Zhu.
During my undergraduate studies, I was fortunate to collaborate with Prof. Gao Huang,
Dr. Jiangmiao Pang
and Prof. Yilin Mo on projects related to reinforcement learning and robotics.
Currently, I am a research intern at the LeCAR Lab at Carnegie Mellon University, advised by Prof. Guanya Shi.
Beyond my research interests, I am also passionate about history, philosophy, economics, and cognitive science.
I've been playing the piano for about ten years and am a big fan of Western classical music, with Beethoven and Mahler being my favorites.
I also enjoy traveling and photography.
I regularly take notes on the papers I’ve read and share my insights. You can find my reading list here.
Feel free to reach out if you are interested in my research or just want to chat!
I'm always open to exploring potential collaborations and engaging in insightful conversations.
My Email: winstongu20[AT]gmail[DOT]com
Google
Scholar  | 
Github  | 
Twitter  
|
|
|
CooHOI: Learning Cooperative Human-Object Interaction with Manipulated Object Dynamics
Jiawei Gao*, Ziqin Wang*, Zeqi Xiao, Jingbo Wang, Tai Wang, Jinkun Cao, Xiaolin Hu, Si Liu, Jifeng Dai, Jiangmiao Pang†
In submission, 2024
[Project Page]
[arXiv]
[BibTeX]
We introduce CooHOI, a framework designed for enabling multiple humanoid characters to learn cooperative human-object interaction tasks,
such as collaboratively carrying large objects.
Initially, each agent independently learns to manipulate the object,
using the object's dynamics as feedback.
Subsequently, CooHOI utilizes these object dynamics as an implicit communication channel to facilitate the coordination learning process between the characters.
|
|
Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Response
Junfeng Long*, Zirui Wang*, Quanyi Li, Jiawei Gao, Liu Cao, Jiangmiao Pang†
2024 International Conference on Learning Representations, ICLR 2024
[Project Page]
[arXiv]
[Code]
[BibTeX]
We present Hybrid Internal Model,
a method enabling the control policy to estimate environmental disturbances
by only explicitly estimating velocity and implicitly simulating the system's response.
|
|
Train Once, Get a Family: State-Adaptive Balances for Offline-to-Online Reinforcement Learning
Shenzhi Wang*, Qisen Yang*, Jiawei Gao, Matthieu Lin, Hao Chen,
Liwei Wu, Ning Jia, Shiji Song, Gao Huang†
2023 Conference on Neural Information Processing Systems, NeurIPS 2023 Spotlight.
[Project Page]
[arXiv]
[Code]
[BibTeX]
We propose FamO2O,
a simple yet effective framework that empowers existing offline-to-online RL algorithms
to determine state-adaptive improvement-constraint balances.
FamO2O utilizes a universal model to train a family of policies with different improvement/constraint intensities,
and a balance model to select a suitable policy for each state.
|
|
OpenDA Project: An open-source platform for coursework and experiences for Tsinghua University undergrads.
[Project Page]
We shared our notes, experiences, insights, and advice from the courses we took during our undergraduate studies.
We hope this can help bridge the "information gap" in undergraduate studies and promote greater educational equity.
|
Recipent for Academic Excellence Scholarship, 2023
Recipent for Outstanding Scientific and Technological Innovation Scholarship, 2023
|
Updated at August. 2024.
Template from Jon Barron.
|
|