Welcome to My Website

About Me

Zhang Zhengming graduated with a Ph.D. in Information and Communication Engineering from the School of Information Science and Engineering, Southeast University in 2023. My main research directions are deep learning, reinforcement learning, federated learning, semi-supervised learning, ultra-high-dimensional data compression and recovery, and wireless communication.

I am passionate about exploring new technologies, such as physical/biological inspired neural network models, and attempting to apply them in all aspects of reality. The future world I envision is one where humans and AI coexist amicably and help each other. At that time, I could chat and discuss work with my AI assistant “Gungun” in a fully self-driving car, while the person I loved was playing games with XR glasses beside me. The AI had already made reasonable travel plans and date locations for us. What a wonderful day!

Work Experience

2023- Present: Senior Engineer at Huawei in Shanghai. As the Owner, I have been conducting research and implementation on wireless ultra-large-scale matrix compression and recovery based on large models. - Design and development of a large MAE model for beam prediction: integrate channel state information and received signal strength information, and build a modal fusion ViT backbone to achieve high-precision classification. - Design and implementation of end-to-end precoding AI algorithm: model the environmental perception, transmission strategy planning, and design of the reception as a symbol-level end-to-end system optimization problem, and develop an AI-based solution to achieve end-to-end optimization. - Design and Implementation of the large white-box AI models for channel matrix compression and recovery: design a multimodal large model based on the white-box Transformer, integrate temporal sequences and frequency domain features, to construct an interpretable and secure AI model. - Do research on channel matrix compression and recovery based on diffusion models: construct a new framework for generating AI models from diffusion models to achieve test time adaptation. - Design a user-friendly and lightweight channel matrix compression and recovery algorithm: adopting the approach of using a small model at the user side and a large model at the network side. ## Personal Skills & Strengths - AI Model Architecture Design: Possesses in-depth expertise in foundational architectures of large models such as Vision Transformer (ViT), Masked Autoencoders (MAE), Vision-Language Models (VLM), and Mixture-of-Experts (MoE). Has a strong research foundation in Transformer architectures and their next-generation variants, including Mamba, White-Box Transformer, and Titan. Previously designed a recurrent Transformer model integrated with GPT that exhibits self-reflective capabilities, outperforming conventional Transformer models in performance. - World Model, Digital Twin, and Agent: Has foundational research experience in world models, digital twins, and intelligent agents. Proposed a dual-loop, three-domain digital twin framework and developed a resource management agent based on this framework to enhance the performance and stability of AI models. - Learning Algorithms: Conducted thorough research on in-context learning and multi-module end-to-end optimization. Introduced context compression, memory, and retrieval algorithms to improve the performance of Transformers applied to data compression. For end-to-end optimization, proposed the LayerLock training strategy to enhance the stability of end-to-end training processes. - Training Frameworks: Has substantial experience in distributed training and federated learning, with familiarity in Distributed Data Parallel (DDP) and mixed-precision training. Previously open-sourced a semi-supervised federated learning distributed training framework: SSFL (https://github.com/jhcknzzm/SSFL-Benchmarking-Semi-supervised-Federated-Learning). - Basic Theories of AI: Holds a strong theoretical background in AI fundamentals, including Loss Landscape, Neural Tangent Kernel (NTK), Explainable AI, and convergence analysis of model training. Capable of analyzing and diagnosing model performance bottlenecks from a theoretical perspective and improving model performance accordingly. ## Publications 1. **Z. Zhang**, et al., "Neurotoxin: Durable Backdoors in Federated Learning," 2022 International Conference on Machine Learning (ICML), 2022. 2. **Z. Zhang**, et al., "Improving Semi-supervised Federated Learning by Reducing the Gradient Diversity of Models," 2021 IEEE International Conference on Big Data (Big Data), 2021, pp. 1214-1225, doi: 10.1109/BigData52589.2021.9671693. 3. P. Ashwinee, A. C. Christopher, **Z. Zhang**, Y. Yang, and M. Prateek. "Teach LLMs to phish: Stealing private information from language models," In: The Twelfth International Conference on Learning Representations. Vienna, Austria, 2024. 4. K. Sun, H. Shi., **Z. Zhang**, Y. Huang, "Ecs-net: Improving weakly supervised semantic segmentation by using connections between class activation maps." In Proceedings of the IEEE/CVF international conference on computer vision (pp. 7283-7292). 5. **Z. Zhang**, Y. Huang, C. Zhang, Q. Zheng, L. Yang and X. You, "Digital Twin-Enhanced Deep Reinforcement Learning for Resource Management in Networks Slicing," in IEEE Transactions on Communications, vol. 72, no. 10, pp. 6209-6224, Oct. 2024, doi: 10.1109/TCOMM.2024.3395698. 6. **Z. Zhang**, T. Ji, H. Shi, C. Li, Y. Huang and L. Yang, "A Self-Supervised Learning-Based Channel Estimation for IRS-Aided Communication Without Ground Truth," in IEEE Transactions on Wireless Communications, vol. 22, no. 8, pp. 5446-5460, Aug. 2023, doi: 10.1109/TWC.2023.3233970. 7. **Z. Zhang**, M. Tian, C. Li, Y. Huang and L. Yang, "Poison Neural Network-Based mmWave Beam Selection and Detoxification With Machine Unlearning," in IEEE Transactions on Communications, vol. 71, no. 2, pp. 877-892, Feb. 2023, doi: 10.1109/TCOMM.2022.3232794. 8. **Z. Zhang**, R. Yang, X. Zhang, C. Li, Y. Huang and L. Yang, "Backdoor Federated Learning-Based mmWave Beam Selection," in IEEE Transactions on Communications, vol. 70, no. 10, pp. 6563-6578, Oct. 2022, doi: 10.1109/TCOMM.2022.3200111. 9. **Z. Zhang**, M. Hua, C. Li, Y. Huang and L. Yang, "Beyond Supervised Power Control in Massive MIMO Network: Simple Deep Neural Network Solutions," in IEEE Transactions on Vehicular Technology, vol. 71, no. 4, pp. 3964-3979, April 2022, doi: 10.1109/TVT.2022.3146434. 10. **Z. Zhang**, H. Chen, M. Hua, C. Li, Y. Huang and L. Yang, "Double Coded Caching in Ultra Dense Networks: Caching and Multicast Scheduling via Deep Reinforcement Learning," in IEEE Transactions on Communications, vol. 68, no. 2, pp. 1071-1086, Feb. 2020, doi: 10.1109/TCOMM.2019.2955490. 11. **Z. Zhang**, Y. Yang, M. Hua, C. Li, Y. Huang and L. Yang, "Proactive Caching for Vehicular Multi-View 3D Video Streaming via Deep Reinforcement Learning," in IEEE Transactions on Wireless Communications, vol. 18, no. 5, pp. 2693-2706, May 2019, doi: 10.1109/TWC.2019.2907077. 12. **Z. Zhang**, M. Hua, C. Li, Y. Huang and L. Yang, "Placement Delivery Array Design via Attention-Based Sequence-to-Sequence Model with Deep Neural Network," in IEEE Wireless Communications Letters, vol. 8, no. 2, pp. 372-375, April 2019, doi: 10.1109/LWC.2018.2873334. 13. S. Xu, **Z. Zhang**, Y. Xu, C. Li and L. Yang, "Deep Reciprocity Calibration for TDD mmWave Massive MIMO Systems Toward 6G," in IEEE Transactions on Wireless Communications, vol. 23, no. 10, pp. 13285-13299, Oct. 2024, doi: 10.1109/TWC.2024.3400616. 14. R. Yang, **Z. Zhang**, X. Zhang, C. Li, Y. Huang and L. Yang, "Meta-Learning for Beam Prediction in a Dual-Band Communication System," in IEEE Transactions on Communications, vol. 71, no. 1, pp. 145-157, Jan. 2023, doi: 10.1109/TCOMM.2022.3223066. 15. M. Tian, **Z. Zhang**, Q. Xu and L. Yang, "A Privacy-Preserved Split Learning Solution for Deep Learning-Based mmWave Beam Selection," in IEEE Communications Letters, doi: 10.1109/LCOMM.2022.3170211. 16. X. Zhang, **Z. Zhang** and L. Yang, "Learning-Based Resource Allocation in Heterogeneous Ultra Dense Network," in IEEE Internet of Things Journal, doi: 10.1109/JIOT.2022.3173210. 17. L. Li, **Z. Zhang** and L. Yang, "Influence of Autoencoder-Based Data Augmentation on Deep Learning-Based Wireless Communication," in IEEE Wireless Communications Letters, vol. 10, no. 9, pp. 2090-2093, Sept. 2021, doi: 10.1109/LWC.2021.3092716. 18. Y. Lin, **Z. Zhang**, Y. Huang, J. Li, F. Shu and L. Hanzo, "Heterogeneous User-Centric Cluster Migration Improves the Connectivity-Handover Trade-Off in Vehicular Networks," in IEEE Transactions on Vehicular Technology, vol. 69, no. 12, pp. 16027-16043, Dec. 2020, doi: 10.1109/TVT.2020.3041521. 19. J. Zhang, S. Xu, **Z. Zhang**, C. Li and L. Yang, "A Denoising Diffusion Probabilistic Model-Based Digital Twinning of ISAC MIMO Channel," in IEEE Internet of Things Journal, vol. 12, no. 15, pp. 29121-29134, 1 Aug.1, 2025, doi: 10.1109/JIOT.2024.3495212. 20. M. Hua, Y. Wang, **Z. Zhang**, C. Li, Y. Huang and L. Yang, "Power-Efficient Communication in UAV-Aided Wireless Sensor Networks," in IEEE Communications Letters, vol. 22, no. 6, pp. 1264-1267, June 2018, doi: 10.1109/LCOMM.2018.2822700. 21. X. Yan, Z. Wang, Y. Jia, **Z. Zhang** and Y. Huang, "Access Point Selection and Beamforming Design for Cell-Free Network: From Fractional Programming to GNN," in IEEE Transactions on Wireless Communications, vol. 23, no. 8, pp. 9345-9360, Aug. 2024, doi: 10.1109/TWC.2024.3361900. 22. P. Zheng, Y. Zhu, Y. Hu, **Z. Zhang** and A. Schmeink, "Federated Learning in Heterogeneous Networks With Unreliable Communication," in IEEE Transactions on Wireless Communications, vol. 23, no. 4, pp. 3823-3838, April 2024, doi: 10.1109/TWC.2023.3311824. 23. K. Sun, W. Lin, H. Shi, Y. Liu, **Z. Zhang**, Y. Huang, "AIRA-DA: Adversarial Image Reconstruction Alignments for Unsupervised Domain Adaptive Object Detection," in IEEE Robotics and Automation Letters, vol. 8, no. 6, pp. 3645-3652, June 2023, doi: 10.1109/LRA.2023.3267692. 24. Y. Huang, C. Xu, C. Zhang, M. Hua and **Z. Zhang**, "An Overview of Intelligent Wireless Communications using Deep Reinforcement Learning," in Journal of Communications and Information Networks, vol. 4, no. 2, pp. 15-29, June 2019, doi: 10.23919/JCIN.2019.8917869. ## Published patents: 1. CN120409543A, Application date: January 31, 2024, Publication Date: August 1, 2025. 2. CN120725084A, Application date: March 28, 2024 Publication date: September 30, 2025. 3. CN120046722A, Application Date: November 24, 2023 Publication Date: May 27, 2025. 4. CN120197733A, Application Date: December 21, 2023. Publication Date: June 24, 2025. 5. WO2025201113A1, Application date: March 18, 2025-Disclosure date: October 2, 2025. 6. WO2025131014A1, Application Date: December 20, 2024-Public Date: June 26, 2025. 7. WO2025108413A1, Application Date: November 22, 2024-Public Date: May 30, 2025. ## International cooperation 1. Collaborated with the teachers and Postdoctoral researchers of Professor Michael W. Mahoney's RISE Lab at the University of UCBerkeley in the United States to conduct research on semi-supervised federated learning algorithms, and the related international conference paper was accepted. 2. Collaborated with the teachers and Ph.D. students of Professor Anke Schmeink's research group from RWTH Aachen University in Germany to conduct research on robust federated learning algorithms in wireless network environments, and the relevant international conference papers were accepted. 3. Collaborated with Ph.D. students from Professor Prateek Mitta's research group at Princeton University in the United States to explore the problem of backdoor attacks in federated learning scenarios. ## Support or Contact - Email: 2352121039@qq.com - [Google Scholar](https://scholar.google.com/citations?user=GePqbSgAAAAJ&hl=en&oi=ao)