Publications

Published Papers

[P1] Zehan Zhu, Yan Huang, Xin Wang, and Jinming Xu, PrivSGP-VR: Differentially Private Variance-Reduced Stochastic Gradient Push with Tight Utility Bounds, Proceedings of the 33rd International Joint Conference on Artificial Intelligence (IJCAI), Jeju, South Korea, Aug. 2024.

[P2] Zehan Zhu, Ye Tian, Yan Huang, Jinming Xu, and Shibo He, R-FAST: Robust Fully-Asynchronous Stochastic Gradient Tracking over General Topology, IEEE Transactions on Signal and Information Processing over Networks, 2024.

[P3] Zehan Zhu, Yan Huang, Chengcheng Zhao, and Jinming Xu, Asynchronous Byzantine-Robust Stochastic Aggregation with Variance Reduction for Distributed Learning, Proceedings of the 62nd IEEE Conference on Decision and Control (CDC), Marina Bay Sands, Singapore, Dec. 2024.

[P4] Changzhi Yan, Zehan Zhu, Youcheng Niu, Cong Wang, Cheng Zhuo, and Jinming Xu, PerfTop: Towards Performance Prediction of Distributed Learning over General Topology, Journal of Parallel and Distributed Computing, 2024.

[P5] Yan Huang, Ying Sun, Zehan Zhu, Changzhi Yan, and Jinming Xu, Tackling Data Heterogeneity: A New Unified Framework for Decentralized SGD with Sample-induced Topology, Proceedings of the 39th International Conference on Machine Learning (ICML), Baltimore, Maryland, USA, Jul. 2022.

Patents

[P3] Zhe Tian, Zehan Zhu, Jinming Xu, Yan Huang, Changzhi Yan, and Xuezhonog Lin, “A privacy protection method for model block aggregation for lithography hotspot detection”, Chinese invention patent, CN116756764B (in Chinese).

[P2] Jinming Xu, Changzhi Yan, Wenchao Meng, Yan Huang, Zehan Zhu, and Youcheng Niu, “Distributed training time prediction method and device for large-scale GPU clusters”, Chinese invention patent, CN116258199B (in Chinese).

[P1] Cheng Zhuo, Xuezhonog Lin, Jinming Xu, Yan Huang, and Zehan Zhu, “Lithography hot spot detection method based on federated personalized learning”, Chinese invention patent, CN113222031A (in Chinese).