This is the homepage of Rui Lu (卢睿) from Tsinghua University. I will be on job market at Oct 2026! Please contact me if you are interested.

Biography

I am a fifth-year Ph.D. student at the Department of Automation at Tsinghua University, advised by Prof. Gao Huang. Prior to that, I received my Bachelor’s degree from Yao Class, the Institute of Interdisciplinary Information Sciences. I was lucky to be advised by Yang Yuan for my undergraduate thesis. I was also a visiting student at Berkeley BAIR Center in 2019 under the supervision of Jitendra Malik and visiting scholar at Princeton University in 2024 supervised by Mengdi Wang.

Research Interest

My research interests are broadly in Machine Learning, Computer Vision, and Reinforcement Learning. In post-GPT era, I am also interested in developing LLMs to discover new knowledge that can give inspiration and insight to human. The main focus and ultimate goal of my research is to develop a principled theory for understanding deep learning’s effectiveness, functionality, and limitations. I am also interested in designing better and faster learning algorithms/architectures.

News

• Our paper Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model? is awarded Best Paper Runner Up at NeurIPS 2025 and Best Paper on AI4Math workshop at ICML 2025!

• Our project How Far is Video Generation from World Model? is released and achieves more than 500k views on X(Twitter).

• Our paper “Efficienttrain: Exploring generalized curriculum learning for training visual backbones” is accepted by ICCV 2023!

• Our paper “Provable General Function Class Representation Learning in Multitask Bandits and MDPs” is accepted as Spotlight in NeuRIPS 2022!

Contact Me

r-lu21 [AT] mails [dot] tsinghua [dot] edu [dot] cn