Han Yu （禹含）

Han Yu （禹含）

PhD Candidate at THU

Hi, I’m a final-year PhD candidate in the Department of Computer Science and Technology at Tsinghua University, advised by Prof. Peng Cui. I got my bachelor degree in the Department of Computer Science and Technology at Tsinghua University in 2021. I also minored in statistics during my undergraduate stage.

My research interest lies in the problem of Out-of-Distribution (OOD) generalization from two complementary aspects. One is designing and implementing algorithms that can generalize to different data distributions, the other is understanding and evaluating the OOD generalization capability of existing algorithms and models.

Recently I have been working on Large Structured-Data Models (LDM), aiming to build a large universal foundation model that is capable of addressing all common types of structured data (tables, time series, graphs, etc.) and various tasks of structured data (classification, regression, missing value imputation, data generation, sample selection, feature selection, etc.). The model can be directly applied to unseen datasets without requiring fine-tuning.

I will graduate in June 2026 and I’m open to both academic and industrial job opportunities. Please feel free to contact me via email if you’re interested.

E-mail: yuh21 [at] mails.tsinghua.edu.cn

Technical Report

  • LimiX: Unleashing Structured-Data Modeling Capability for Generalist Intelligence, Technical Report, 2025, [arxiv] [code]
    • Team LimiX (core contributor)

Conference Proceedings

  • Generating Risky Samples with Conformity Constraints via Diffusion Models, AAAI 2026
    • Han Yu, Hao Zou, Xingxuan Zhang, Zhengyi Wang, Yue He, Kehan Li, Peng Cui
  • Error Slice Discovery via Manifold Compactness, AAAI 2026 [arxiv]
    • Han Yu, Hao Zou, Jiashuo Liu, Renzhe Xu, Yue He, Xingxuan Zhang, Peng Cui
  • ODP-Bench: Benchmarking Out-of-Distribution Performance Prediction, ICCV 2025 [paper] [arxiv]
    • Han Yu*, Kehan Li*, Dongbai Li, Yue He, Xingxuan Zhang, Peng Cui
  • PDMC: Generating Feasible Algorithmic Recourse via Perturbation Data Manifold Constraint, KDD 2025 [paper]
    • Zimu Wang, Hao Zou, Han Yu, Shaohua Fan, Haotian Wang, Yue He, Peng Cui
  • Generalization of Transformers with In-Context Learning: An Empirical Study, ICLR 2025 [paper]
    • Xingxuan Zhang*, Haoran Wang*, Jiansheng Li, Yuan Xue, Shikai Guan, Renzhe Xu, Hao Zou, Han Yu, Peng Cui
  • Quantization Meets OOD: Generalizable Quantization-aware Training from a Flatness Perspective, MM 2025 [arxiv]
    • Jiacheng Jiang, Yuan Meng, Chen Tang, Han Yu, Qun Li, Zhi Wang, Wenwu Zhu
  • Domain-wise Data Acquisition to Improve Performance under Distribution Shift, ICML 2024 [paper] [code]
    • Yue He*, Dongbai Li*, Pengfei Tian, Han Yu, Jiashuo Liu, Hao Zou, Peng Cui
  • Rethinking the Evaluation Protocol of Domain Generalization, CVPR 2024 [paper] [arxiv] [code]
    • Han Yu, Xingxuan Zhang, Renzhe Xu, Jiashuo Liu, Yue He, Peng Cui
  • Flatness-Aware Minimization for Domain Generalization, ICCV 2023 [paper] [arxiv]
    • Xingxuan Zhang, Renzhe Xu, Han Yu, Yancheng Dong, Pengfei Tian, Peng Cui
  • NICO++: Towards Better Benchmarking for Domain Generalization, CVPR 2023 [paper] [arxiv] [code]
    • Xingxuan Zhang*, Yue He*, Renzhe Xu, Han Yu, Zheyan Shen, Peng Cui
  • Gradient Norm Aware Minimization Seeks First-Order Flatness and Improves Generalization, CVPR 2023, (Highlight, top 2.5%) [paper] [arxiv] [code]
    • Xingxuan Zhang*, Renzhe Xu*, Han Yu, Hao Zou, Peng Cui
  • Stable Learning via Sparse Variable Independence, AAAI 2023 (Oral) [paper] [arxiv] [code]
    • Han Yu, Peng Cui, Yue He, Zheyan Shen, Yong Lin, Renzhe Xu, Xingxuan Zhang
  • NICO Challenge: Out-of-Distribution Generalization for Image Recognition Challenges, ECCV 2022 workshop, [paper] [workshop]
    • Xingxuan Zhang, Yue He, Tan Wang, Jiaxin Qi, Han Yu, Zimu Wang, Jie Peng, Renzhe Xu, Zheyan Shen, Yulei Niu, Hanwang Zhang, Peng Cui

Preprints

  • Sample Weight Averaging for Stable Prediction, [arxiv]
    • Han Yu, Yue He, Renzhe Xu, Dongbai Li, Jiayin Zhang, Wenchao Zou, Peng Cui
  • A Survey on Evaluation of Out-of-Distribution Generalization, [arxiv]
    • Han Yu, Jiashuo Liu, Xingxuan Zhang, Jiayun Wu, Peng Cui
  • Meta Adaptive Task Sampling for Few-Domain Generalization, [arxiv]
    • Zheyan Shen*, Han Yu*, Peng Cui, Jiashuo Liu, Xingxuan Zhang, Linjun Zhou, Furui Liu
  • Towards Out-of-Distribution Generalization: A Survey, [arxiv]
    • Jiashuo Liu*, Zheyan Shen*, Yue He, Xingxuan Zhang, Renzhe Xu, Han Yu, Peng Cui

* indicates equal contributions.

Professional Activities

  • Top reviewer: NeurIPS 2022 (10%), KDD 2025 (10%~20%).

  • Conference reviewer / Program committee member
    • ICML, NeurIPS, ICLR, CVPR, ICCV, KDD, AAAI, MM, WACV, AISTATS, UAI, ACML.
  • Journal reviewer
    • International Journal of Computer Vision (IJCV).
    • IEEE Transactions on Multimedia (TMM).
    • IEEE Transactions on Neural Networks and Learning Systems (TNNLS).
    • ACM Transactions on Intelligent Systems and Technology (TIST).

Miscellaneous

During my free time, I’m enthusiastic about playing table tennis and was the captain of Tsinghua CS table tennis team. I can play well with both double-sided inverted rubber and a long pips style. I also enjoy watching Japanese cartoons.