Personalized Federted Learning

In the standard Federated Learning (FL) paradigm, multiple clients aim to learn models cooperatively without directly sharing their private data. However, clients may have different data distributions and training objectives that cannot be customized in FL. As a result, the performance of the learned model is downgraded, and FL learning cannot fully mine the data of clients, especially clients with non-iid data. My research explores the possibility of personalized FL architecture and FL aggregation algorithms to achieve personalized learning objectives for each client and minimize/remove the performance sacrifice of any other client. The research progress and outcomes are distributed in the paper collection.

  • G. Yuan, J. Li, Y. Huang, Z. Xie, J. Pang, Z. Cai
    Independence and Unity: Unseen Domain Segmentation Based on Federated Learning
    IEEE Internet of Things Journal, 2023. (IF: 10.6)
  • C. Jing, Y. Huang, Y. Zhuang, L. Sun, Y. Huang, Z. Xiao, X. Ding
    Exploring Personalization via Federated Representation Learning on Non-IID Data
    Neural Networks, 2023. (IF: 9.657)
  • Z. Xie, Y. Huang, D. Yu, R.M. Parizi, Y. Zheng, J. Pang
    FedEE: A Federated Graph Learning Solution for Extended Enterprise Collaboration
    IEEE Transactions on Industrial Informatics, 2022. (IF: 11.648)
  • J. Pang, Y. Huang, Z. Xie, J. Li, Z. Cai
    Collaborative city digital twin for the COVID-19 pandemic: A federated learning solution
    Tsinghua science and technology, 2021.
    (Excellent Paper Award)(Hot Paper by Web of Science)
  • J. Pang, Y. Huang*, Z. Xie, Q. Han, Z. Cai
    Realizing the Heterogeneity: A Self-Organized Federated Learning Framework for IoT IEEE Internet of Things Journal, 2020. (IF: 10.6)

VR Cybersecurity Education for Middle School Students

This is an NSA/NSF GenCyber program supported project. We build six modules: Digital Footprints, Trojan Horse/Ransomware, Cryptography, Hacking Ethics, Authentication & Authorization, and Phishing to augment cybersecurity education for middle school students using VR lectures and games. A video of this project will be posted when it is fully completed in Summer 2024.

Cybersecurity Park