Beijing University of Posts and Telecommunications
School of Computer Science · Computer Science and Technology · B.Eng.
Rank: 30 / 396 (Top 7.6%)
About Me
I am a Ph.D. student in Software Engineering at the School of Computer Science and Engineering, Beihang University, advised by Prof. Wenfei Fan and Assoc. Prof. Ping Lu.
My research primarily centers on Data-Centric AI, with a focus on Data Quality, Knowledge Discovery, and Model Reliability in graph learning. My work includes LLM-assisted feature extraction and ranking, graph data quality enhancement, graph data mining, and GNN explainability.
- Beyond Research
Education & Work Experience
Education
Beihang University
School of Computer Science and Engineering · Software Engineering · Ph.D. Student
Research: Graph Data Quality · Graph Data Mining · Graph Model Reliability
News
CMC First Prize MCM Honorable Mention CMC First Prize Innovation Program Municipal Project Award TODS · J. Accepted TODS · J. Accepted KDD · Conf. Accepted ICDE · Conf. Accepted SIGMOD · Conf. Accepted ICDE · Conf. Accepted KDD · Conf. AcceptedWork Experience
Shenzhen Institute of Computing Sciences Research Intern
Conducted research on graph data governance and enhancement, GNN explainability, graph computing systems, and low-resource adaptation of LLMs in large-scale data settings. This work resulted in multiple publications at CCF-A venues, including ACM TODS and IEEE ICDE.
Meituan Algorithm Research Intern · Core Local Commerce / M17
Contributed to web-data cleaning and quality assessment for foundation-model pretraining; deployed a local 7B model and performed full-parameter SFT. The fine-tuned model achieved approximately 96% of the GPT-4 baseline in ROUGE-L and outperformed conventional methods by approximately 6%.
Publications
Peer-reviewed and formally published or accepted work. Entries follow the official bibliographic record.
2026
- ICDE 2026 CCF-A ConferenceExplaining GNN Negatives Globally and Locally
2026 IEEE 42nd International Conference on Data Engineering, pp. 1603–1616.
DOI: 10.1109/ICDE65706.2026.00123 - KDD 2026 CCF-A ConferenceAccelerating Influence Function Estimation for Large Language Models: A Practical Design
Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Vol. 2, pp. 5927–5938.
DOI: 10.1145/3770855.3817953
2025
- ICDE 2025 CCF-A ConferenceImputing Sparse and Noisy Labels for GNNs
2025 IEEE 41st International Conference on Data Engineering, pp. 2295–2308.
DOI: 10.1109/ICDE65448.2025.00174 - SIGMOD 2025 CCF-A ConferenceRule-Based Graph Cleaning with GPUs on a Single Machine
Proceedings of the ACM on Management of Data 3(3), Article 166, pp. 166:1–166:27.
DOI: 10.1145/3725303
2024
- ACM TODS CCF-A JournalMaking It Tractable to Detect and Correct Errors in Graphs
ACM Transactions on Database Systems 49(4), Article 16, pp. 16:1–16:75.
DOI: 10.1145/3702315 - KDD 2024 CCF-A ConferenceEfficient Mixture of Experts based on Large Language Models for Low-Resource Data Preprocessing
Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp. 3690–3701.
DOI: 10.1145/3637528.3671873 - ACM TODS CCF-A JournalLinking Entities across Relations and Graphs
ACM Transactions on Database Systems 49(1), Article 2, pp. 2:1–2:50.
DOI: 10.1145/3639363
Manuscripts / Under Review
- Under Revision
OpenMind: Query-Time Diversification via Logical Rules
- Under Review
ComCE: Community-Aware Counterfactual Explanations for Graph Neural Network
Honors & Awards
- First Prize, Beijing Undergraduate Mathematics Competition, 2019 & 2020
- First Prize, Chinese Mathematics Competitions, 2019 & 2020
- Undergraduate Scholarship, Beijing University of Posts and Telecommunications, 2019–2021
- Academic Scholarship, School of Computer Science, Beihang University, 2022–2025
- Honorable Mention, Mathematical Contest in Modeling, 2020
- Beijing municipal-level Undergraduate Innovation and Entrepreneurship Training Program award, 2020
Academic Service
External Reviewer, IEEE ICDE, 2024–2026
Patent
A Method, Apparatus, Device, and Medium for Graph Data Augmentation
Chinese invention patent application, 2025
Contact
For research discussions and professional opportunities, please get in touch by email.
Email: pangkehan@buaa.edu.cn
Alternative: k3hanpang@gmail.com
Address: Beihang UniversityNo. 37 Xueyuan RoadHaidian District, Beijing, China