Changfan Pan (born January 26, 1998) is a researcher specializing in causal inference, synthetic data generation, and federated learning, particularly in the education domain. He is currently pursuing a Master of Engineering in Computer Science at Zhejiang Normal University, China, where he is researching interpretable models and dataset generation methods. He previously earned a Bachelor’s degree in Telecommunication Engineering from Beijing University of Posts and Telecommunications. His work experience includes roles at Zhongxing Telecom Equipment, China Mobile Communications Group, and Ericsson, focusing on 5G networking, wireless network maintenance, and backend development. He has published in top conferences such as AAAI and IEEE ICDM. He has received numerous honors, including a Kaggle Bronze Medal and a First-Class Scholarship. His technical skills include Python, C++, Pytorch, and Docker, and he is proficient in Mandarin and English. ππ¬π»
Profile
Education π
Changfan Pan is currently pursuing an M.E. in Computer Science at Zhejiang Normal University (2022-2025) with a GPA of 3.70/5. His master’s thesis focuses on interpretable models and dataset generation based on causal inference under the supervision of Prof. Jia Zhu. He has completed coursework in Artificial Intelligence (91), Data Mining (88), and Advanced Database Technology (86). Previously, he earned a B.E. in Telecommunication Engineering from Beijing University of Posts and Telecommunications (2016-2020), graduating with a GPA of 84/100. His bachelorβs thesis explored viewpoint prediction algorithms in VR, supervised by Prof. Yitong Liu. His coursework included Probability Theory (92), Programming Practices (87), and Pattern Recognition (88). He has developed strong expertise in AI, data science, and federated learning, positioning him for impactful contributions to academia and industry. πππ‘
Experience π¨βπ«
Changfan Pan has extensive experience in telecommunications and AI research. He interned at Zhongxing Telecom Equipment (2019), setting up 5G signal transceivers and deploying co-located base stations. At China Mobile (2020-2021), he maintained 2G/4G/5G networks, optimized signal coverage, and managed mobile cloud services. At Ericsson (2022), he developed backend servers using 5G protocols and designed performance test cases. His research includes developing a VR electromagnetic wave teaching platform, designing intelligent teaching assessment tools for rural education, and implementing federated knowledge graph completion systems. His technical expertise spans backend development, federated learning, and AI model interpretability. π‘ππ€
Research Interests π¬
Changfan Panβs research focuses on causal inference, synthetic data generation, and federated learning. He explores interpretable methods in knowledge tracing and stable attribution using local surrogate models. His work on synthetic data generation emphasizes fairness and privacy preservation. In federated learning, he designs decentralized frameworks for education applications, enhancing privacy-preserving knowledge graph completion. His published research includes papers in AAAI and IEEE ICDM, covering cross-domain multi-modality classification and trustworthy AI mechanisms. His projects include VR-based educational tools and intelligent assessment systems for rural teachers. His research aims to bridge AI advancements with practical applications in education and data privacy. πππ
Awards & Recognitions π
Changfan Pan has received multiple accolades for his contributions to AI and telecommunications. He was awarded the National Undergraduate Innovation and Entrepreneurship Training Program honor (2019). At China Mobile, he won First Prize for Wireless Network Maintenance Quality (2022). In AI competitions, he achieved a Bronze Medal (Top 6%) in a Kaggle Featured Code Competition (2023). Academically, he received the First-Class Scholarship at Zhejiang Normal University (2023). His consistent excellence in research, industry projects, and competitive AI challenges highlight his strong analytical and problem-solving skills. π
πποΈ
Publications π
1. Changfan Pan, Jia Zhu, Qing Wang, Stable Attribution with Local Linear Surrogate Model, ThirtyEighth AAAIΒ Β Β Β Β Β Conference on Artificial Intelligence. (CCF-A)
2. Qing Wang, Jia Zhu, Changfan Pan, Yilong Ji, and Hanghui Guo, Dual trus