Runtian Zhai

翟润天

PhD Student

Third-year PhD
Machine Learning
Computer Science Department (CSD)
School of Computer Science (SCS)
Carnegie Mellon University (CMU)

Email: rzhai at cmu dot edu
Office: GHC 5105

Photo

Third-year PhD
Machine Learning
Computer Science Department (CSD)
School of Computer Science (SCS)
Carnegie Mellon University (CMU)

Email: rzhai at cmu dot edu
Office: GHC 5105

Bio [CV]
I am a third-year PhD student of CMU SCS Computer Science Department, co-advised by Zico Kolter and Pradeep Ravikumar. My research interest is the fundamental theory of learning, and furthermore how to close the gap between theory and practice. I am most interested in generalization theory, and recently I am mainly focusing on analyzing the generalization of representation learning, especially with big models, with kernel and spectral graph theory. I am also studying OOD generalization, which is how a model can generalize to a test data distribution different from the training distribution, and related topics include domain adaptaion, continual learning, algorithmic fairness, etc.
I received my Bachelor's degree in computer science and applied math (double degree) from Peking University. As an undergraduate I was advised by Liwei Wang. I also visited UCLA in the summer of 2019 and worked with Cho-Jui Hsieh. In the summer of 2022, I worked at Amazon Alexa AI at Sunnyvale as an applied scientist intern. From Sept 2019 to Jun 2020 I worked as a full-time research intern in Microsoft Research Asia (MSRA) machine learning group at Beijing.
Service
Peer review:
  • ICML 2022, 2023
  • ICCV 2023
  • KDD 2023
  • AISTATS 2023
  • ICLR 2023
  • NeurIPS 2022, 2023
  • JMLR
  • NeurIPS workshops: ML Safety'22, TSRML'22
  • ICML workshop: PODS'22
Teaching:
  • CMU 10-701: Introduction to Machine Learning Fall 2022 (Head TA)
News
  • Two papers at ICLR 2023 workshops.
  • One paper accepted by ICLR 2023.
  • One paper accepted by NeurIPS 2021.
  • Arrived in Pittsburgh. (8/26/21)
  • One paper accepted by ICML 2021.
  • Graduated from PKU. (Picture)
Links