About Me

• I will start my doctoral research with the Statistics and Machine Learning Laboratory at Graduate School of Informatics, Kyoto University, supervised by Prof. Junya Honda.

• My current research interests are mainly in learning theory and online optimization.

• I received my Master’s Degree in Computer Technology, at School of Computer Science and Technology, Harbin Institute of Technology (Shenzhen), supervised by Prof. Qing Liao.

• I received my Bachelor’s Degree in Computer Science and Technology, at School of Computer Science, SWPU in 2022.

Research Area

• Learning Theory
• Online Optimization
• Data Mining

Work experience

• Algorithm Intern (2024.04-2024.07), Tencent, PCG Department, QQ short video recommendation. I’m mainly responsible for the multi-objective fusion (MTF) of precision ranking scores based on implicit Q-learning (IQL) reinforcement learning algorithm.

Manuscripts and Publications

• Zhang H, Zheng D, Yang X, et al. FedDCSR: Federated Cross-domain Sequential Recommendation via Disentangled Representation Learning[C]. SIAM International Conference on Data Mining, 2024. [paper] [code]

• Zhang H, Zheng D, Zhong L, et al. FedHCDR: Federated Cross-Domain Recommendation with Hypergraph Signal Decoupling[C]. ECML-PKDD, 2024. [paper] [code]

• Zheng D, Zhang H, Zhai J, et al. FedCSR: A Federated Framework for Multi-Platform Cross-Domain Sequential Recommendation with Dual Contrastive Learning[C]. COLING, 2025. [paper] [code]

Software or Toolkit

• FedCom: A new method of multi-task clustered federated learning. In this method, we perform dynamic partition of clusters based on community detection, which can alleviate the negative impact caused by the early-stage partition errors.

• FedAO: A toolbox for federated learning, aiming to provide implementations of FedAvg, FedProx, Ditto, etc. in multiple versions, such as Pytorch/Tensorflow, single-machine/distributed, synchronized/asynchronous and so on.

• Numerical Analysis: A project for the implementation of the algorithm in the numerical analysis (by Timothy Sauer) using Python+Numpy+Pytorch.

Selected Award

• 2021 First Prize of China College Students Computer Design Competition [Cert]
• 2020 Golden Medal of China Parallel Application Challenge On Domestic CPU [Cert]
• 2020 Second Prize of ”Blue Bridge Cup” National Software Competition Final [Cert]
• 2019 First Prize of National English Competition for College Students (NECCS) [Cert]

Social Service

• Former captain of the student team SCPCLab
• Volunteer for the 3rd Sichuan Robot Competition [Cert]
• 2020 Class Tutor for Computer Science and Technology major at SWPU [Cert]

Selected Blog Post

• 谱图论：Laplacian算子及其谱性质
• 迁移学习：互信息的变分上下界
• SICP：惰性求值、流和尾递归（Python实现）
• 分布式机器学习：异步SGD和Hogwild!算法（Pytorch）
• 随机算法：蒙特卡洛和拉斯维加斯算法
• 因果推断：解决推荐系统公平性的新思路

Folks

• include-yy :

  • 用 ReScript 写个神经网络 hello world

  • 什么是 dsl 和作为 dsl 的 cl-loop

• Koshkaaa :
  • 为什么go和rust语言都舍弃了继承？
  • LLVM 参考链接
• TTS-S :
  • 小样本学习One-shot
  • 置换环
• Bruce :
  • Convex Hull - Graham scan 详解
  • 优化1: 求解顶点覆盖问题和度量斯坦纳树问题的近似算法
• TheWitcher :
  • 机器学习理论、算法理论与信息通信理论资源推荐
• 白小鱼 :
  • [科研小记] 联邦学习算法收敛性分析 Convergence of Federated Learning
  • 联邦学习开源框架（2022.07.12）
• lokinko :
  • Machine Unlearning | 如何看待机器学习的数据遗忘？
  • 「联邦学习」— Personalization 技术