Undergraduate at College of Engineering, Peking UniversityI am a fourth year undergraduate student at the College of Engineering, Peking University. I expect to receive my bachelor's degree in Robotics Engineering in June 2026. My GPA is 3.86/4.0, ranking 1st/26 in the Robotics major and top 5% (156 students) in the College of Engineering.
I am very fortunate to be advised by Prof. Yisen Wang in self-supervised learning, and Prof. Pengcheng You in reinforcement learning. And I am currently a visiting research student at Electrical and Computer Engineering, Johns Hopkins University, advised by Prof. Enrique Mallada in data-driven learning and control in nonlinear systems (especially in Hamiltonian systems).
My research interest covers self-supervised learning, reinforcement learning, and data-driven learning methods in Hamiltonian systems. I am currently interested in generalizing transfer learning paradigm to safe reinforcement learning settings, utilizing offline dataset from source environment to enhance sample efficiency in target environment. I am actively applying for a PhD position for Fall 2026!
Academic Transcript
Education
-
Peking UniversityB.S. in Robotics EngineeringSep. 2022 - Jun. 2026
Honors & Awards
-
Award for Academic Innovation, College of Engineering, Peking University (top 1%)2025
-
China National Scholarship2025
-
Beijing Natural Science Foundation Start-up Fellowship (50,000 CNY)2025
-
Merit Student Pacesetter, Peking University2025
-
Peking University Students of the Year Finalist (20 Candidates, Most Prestigious Award for All University Students)2024
-
China National Scholarship2024
-
Merit Student Pacesetter, Peking University2024
-
The First Prize in the Final Round of Chinese National College Mathematics Competition (Top 20 among all Participants)2024
-
The First Prize in Preliminary Round of Chinese National College Mathematics Competition2023
-
First Prize in Chinese National College Physics Competition2023
Experience
-
Johns Hopkins UniversityVisiting StudentOct. 2025 - Feb. 2026
News
Selected Publications (view all )
Projection Head is Secretly an Information Bottleneck
Zhuo Ouyang, Kaiwen Hu, Qi Zhang, Yifei Wang, Yisen Wang
International Conference on Learning Representations (ICLR) 2025
We develop a new theoretical understanding for the role of the projection head in contrastive learning from the information-theoretic perspective. Mathematically, we rigorously derive both lower and upper bounds for the downstream performance of the features preceding the projection head.
Our findings indicate theoretical principles for designing an effective projection head: it should act as an information bottleneck, filtering out the irrelevant information and preserving the essential information for the contrastive objective.
Based on theoretical principles, we propose two categories of methods to improve projection head design, namely training regularization and structural regularization, that outperform previous approaches across a range of datasets and contrastive methods.
Projection Head is Secretly an Information Bottleneck
Zhuo Ouyang, Kaiwen Hu, Qi Zhang, Yifei Wang, Yisen Wang
International Conference on Learning Representations (ICLR) 2025
We develop a new theoretical understanding for the role of the projection head in contrastive learning from the information-theoretic perspective. Mathematically, we rigorously derive both lower and upper bounds for the downstream performance of the features preceding the projection head.
Our findings indicate theoretical principles for designing an effective projection head: it should act as an information bottleneck, filtering out the irrelevant information and preserving the essential information for the contrastive objective.
Based on theoretical principles, we propose two categories of methods to improve projection head design, namely training regularization and structural regularization, that outperform previous approaches across a range of datasets and contrastive methods.