Jiaye Teng

alt text 

Jiaye Teng
Email: tjy20 [at] mails [dot] tsinghua [dot] edu [dot] cn
Research interests: generalization theory, machine learning and statistics

About Me

I am now a Ph.D student at Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University. I am very fortunate to have Professor Yang Yuan as my advisor. I am now visiting Princeton University, advised by Professor Sanjeev Arora. Previously, I did my undergraduate at Shanghai University of Finance and Economics, advised by Professor Yang Bai.

Research Interests

My research interests lie in the machine learning theory, especially in the generalization of neural networks. I am also interested in statistical learning theory, including conformal inference and causal inference.
If you want to have a chat, always feel free to contact me through Email!

News

  • 1 paper accepted in NeurIPS, Sep 2023
  • Going to visit Princeton from Aug 2023!
  • 2 papers accepted in ICLR, Jan 2023
  • 2 papers accepted in ICML, May 2023
  • Publications

    • When do Models Generalize? A Perspective from Data-Algorithm Compatibility [arxiv]
      Jing Xu* , Jiaye Teng* , Yang Yuan, Andrew Chi-Chih Yao
      Conference on Neural Information Processing Systems (NeurIPS) 2023

    • Finding Generalization Measures by Contrasting Signal and Noise [paper]
      Jiaye Teng*, Bohang Zhang*, Ruichen Li*, Haowei He*, Yequan Wang, Yan Tian, Yang Yuan International Conference on Machine Learning (ICML) 2023

    • On Uni-Modal Feature Learning in Supervised Multi-Modal Learning [arxiv]
      Chenzhuang Du*, Jiaye Teng*, Tingle Li, Yichen Liu, Tianyuan Yuan, Yue Wang, Yang Yuan, Hang Zhao
      International Conference on Machine Learning (ICML) 2023

    • Predictive Inference with Feature Conformal Prediction [arxiv]
      Jiaye Teng*, Chuan Wen*, Dinghuai Zhang*, Yoshua Bengio, Yang Gao, Yang Yuan
      International Conference on Learning Representations (ICLR) 2023

    • Benign Overfitting in Classification: Provably Counter Label Noise with Larger Models [arxiv]
      Kaiyue Wen*, Jiaye Teng*, Jingzhao Zhang
      International Conference on Learning Representations (ICLR) 2023

    • Fighting Fire with Fire: Avoiding DNN Shortcuts through Priming [arxiv] [paper]
      Chuan Wen, Jianing Qian, Jierui Lin, Jiaye Teng, Dinesh Jayaraman, Yang Gao
      International Conference on Machine Learning (ICML) 2022

    • Towards Understanding Generalization via Decomposing Excess Risk Dynamics [arxiv] [paper]
      Jiaye Teng* , Jianhao Ma* , Yang Yuan
      International Conference on Learning Representations (ICLR) 2022

    • Can Pretext-Based Self-Supervised Learning Be Boosted by Downstream Data? A Theoretical Analysis [arxiv]
      Jiaye Teng* , Weiran Huang* , Haowei He*
      International Conference on Artificial Intelligence and Statistics (AISTATS) 2022

    • T-SCI: A Two-Stage Conformal Inference Algorithm with Guaranteed Coverage for Cox-MLP [arxiv] [paper]
      Jiaye Teng* , Zeren Tan* , Yang Yuan
      International Conference on Machine Learning (ICML) 2021

    Academic Service

    Honors

    • Top Reviewer in NeurIPS 2022
    • 2022 National Scholarship at Tsinghua University