Publications

publications by categories in reversed chronological order.

Preprint

  1. Preprint
    SAC-Copula: Quality-Preserving Watermarking for Diffusion Language Models via Smooth Correlated Gumbel Fields
    Baixin Li, and Haiyun He
    Submitted, May 2026
  2. Preprint
    Fundamental Trade-Offs in Multi-Bit Watermarking of Stochastic Processes
    Haiyun He, Yepeng Liu, Zhuoer Shen, Ziqiao Wang, Yongyi Mao, and Yuheng Bu
    Submitted, May 2026
    arXiv

Journal

  1. Journal
    Information-Theoretic Generalization Bounds for Deep Neural Networks
    Haiyun He, and Ziv Goldfeld
    IEEE Transactions on Information Theory, May 2025
    arXiv HTML
  2. Journal
    Information-Theoretic Characterization of the Generalization Error for Iterative Semi-Supervised Learning
    Haiyun He, Hanshu Yan, and Vincent Y. F. Tan
    Journal of Machine Learning Research, Aug 2022
    arXiv HTML
  3. Journal
    Optimal Change-Point Detection with Training Sequences in the Large and Moderate Deviations Regimes
    Haiyun He, Qiaosheng (Eric) Zhang, and Vincent Y. F. Tan
    IEEE Transactions on Information Theory, Oct 2021
    arXiv HTML
  4. Journal
    Distributed Detection with Empirically Observed Statistics
    Haiyun He, Lin Zhou, and Vincent Y. F. Tan
    IEEE Transactions on Information Theory, Jul 2020
    arXiv HTML
  5. Journal
    Joint altitude and beamwidth optimization for UAV-enabled multiuser communications
    Haiyun He, Shuowen Zhang, Yong Zeng, and Rui Zhang
    IEEE Communications Letters, Feb 2018
    arXiv HTML

CS Conf.

  1. CS Conf.
    PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks
    Zhenxin Ai, and Haiyun He
    ICML, May 2026
    arXiv
  2. CS Conf.
    Theoretically Grounded Framework for LLM Watermarking: A Distribution-Adaptive Approach
    Haiyun He*, Yepeng Liu*, Ziqiao Wang, Yongyi Mao, and Yuheng Bu
    NeurIPS, Dec 2025 (Presented at 1st Workshop on GenAI Watermarking (WMARK@ICLR))
    arXiv
  3. CS Conf.
    Information-Theoretic Generalization Bounds for Deep Neural Networks
    Haiyun He, Christina Lee Yu, and Ziv Goldfeld
    In InfoCog Workshop @ NeurIPS  (*Contributed Talk*), Dec 2023
    HTML Slides
  4. CS Conf.
    How Does Pseudo-Labeling Affect the Generalization Error of the Semi-Supervised Gibbs Algorithm?
    Haiyun He, Gholamali Aminian, Yuheng Bu, Miguel Rodrigues, and Vincent Y. F. Tan
    In International Conference on Artificial Intelligence and Statistics (AISTATS), 2023
    arXiv HTML

IT Conf.

  1. IT Conf.
    On the Generalization of Knowledge Distillation: An Information-Theoretic View
    Bingying Li, and Haiyun He
    In IEEE International Symposium on Information Theory (ISIT), 2026
    arXiv
  2. IT Conf.
    Distributional Information Embedding: A Framework for Multi-bit Watermarking
    Haiyun He, Yepeng Liu, Ziqiao Wang, Yongyi Mao, and Yuheng Bu
    Asia Pacific Workshop on Data Science and Information Theory (APWDSIT), Oct 2025 (Presented at 1st Workshop on GenAI Watermarking (WMARK@ICLR))
    arXiv
  3. IT Conf.
    Hierarchical Generalization Bounds for Deep Neural Networks
    Haiyun He, Christina Lee Yu, and Ziv Goldfeld
    In IEEE International Symposium on Information Theory (ISIT), Oct 2024
    HTML
  4. IT Conf.
    Optimal Resolution of Change-Point Detection with Empirically Observed Statistics and Erasures
    Haiyun He, Qiaosheng (Eric) Zhang, and Vincent Y. F. Tan
    In International Symposium on Information Theory and Its Applications (ISITA), Oct 2020
    HTML
  5. IT Conf.
    Distributed Detection with Empirically Observed Statistics
    Haiyun He, Lin Zhou, and Vincent Y. F. Tan
    In IEEE Information Theory Workshop (ITW), Oct 2019
    HTML