Hi there! My name is Yu Lei(宇 雷), while [宇] means universe and [雷] means thunder in Chinese.
I am a senior undergraduate student from Shanghai Jiao Tong University, majoring in Electronic Science and Technology.
Currently, I am a research intern at K-Lab at Carnegie Mellon University, advised by Prof. Kris Kitani, working on humanoid whole-body control, efficient world model and intersection in between.
I was also fortunated to work with Prof. Zhuowen Tu at CogSci Department of UCSD in 2023. At Shanghai Jiao Tong University, I started to learn the beautiful artificial intelligence under the guidance of Prof. Bingbing Ni and collaborate with Prof. Zhijie Deng.
I am actively looking for PhD opportunities in 2025 Fall!
Github |
Google Scholar |
X
Email: KnightYU2002 [AT] gmail.com / yulei3 [AT] andrew.cmu.edu
Research Interests
I'm especially interested in scientifically understanding and building human-like intelligence that can percept, reason and interact with the physical world. My ultimate goal is to build embodied intelligent systems that can help us explore physical world and expand our boundaries of capability perceptually, intellectually and physically. To achieve this, I think there are several key efforts we should make:
- Generative Intelligence: how can we use generative models as mental simulation engines to support learning and planning? And how can we build generative models that can effectively extrapolate from training data?
- (Internal) World Model for Agents: How should an embodied agent represent the dynamic environment around it? It is crucial to learn an internal world model for agents because the model reflects agents' understanding of states and causality in physical world.
- Self-Improving Intelligence: How can we enable embodied agents to autonomously improve and correct themselves, with ability of learning to learn, learning to plan, and planning to learn?
Publications and Preprints
* indicates equal contribution
Humanoid Whole-body Control via Large Neural Dynamics Model
Manuscript in Progress
Advancing Text-to-3D Generation with Linearized Lookahead Variational Score Distillation
ICLR 2025(Under Review)
Bayesian Diffusion Models for 3D Shape Reconstruction
CVPR 2024 [Paper] [Code] [Project page]
Bayesian Exploration of Pre-trained Models for Low-shot Image Classification
CVPR 2024 [Paper]
FocalDreamer: Text-driven 3D Editing via Focal-fusion Assembly
AAAI 2024 [Paper] [Code] [Project page]
Side Projects
Investigation in Fixed-point Problem in Video World Model
Personal research exploration [Blog(building)]
My major is relatively far from what I am doing right now. What have I done during my undergrad years?
Analysis of Active Reflector Surface Shape Adjustment for FAST Based on Iterative Optimization
Third prize of CUMCM 2021 [Paper(in Chinese)]
2D Pure Azimuth Passive UAV Localization Based on Approximate Iterative Algorithms
Second prize of CUMCM 2022 [Paper(in Chinese)]
Design of Active Two-Way Audio Amplifier Circuit
First prize of TI Cup NUEDC 2022
Frequency Modulation Demodulation Circuit Based on Phase-Locked Loop Technology
Course Project of High Frequency Electronic Circuit [Report(in Chinese)]
Honors and Awards
- First prize of TI Cup National Undergraduate Students Electronic Design Competition, 2022
- First prize of China College Students Mathematics Competiton, 2021
- Second prize of China Undergraduate Mathematical Contest in Modeling, 2022
- Huawei Scholarship in Shanghai Jiao Tong University, 2022
- Excellent Scholarships for Undergraduate Students in Shanghai Jiao Tong University, 2023,2022,2021
Reviewer Services
- Conference Reviewer:
- Machine Learning: NeurIPS 2024, ICLR 2025
Experience
 |
Carnegie Mellon University
2024.10 - 2025.04 Visiting Researcher Research Advisor: Prof. Kris Kitani |
 |
University of California, San Diego
2023.08 - 2023.12 Research Intern Research Advisor: Prof. Zhuowen Tu |
 |
Shanghai Jiao Tong University
2020.09 - 2025.06 Undergraduate Student |