Hi, there! Welcome to Zhengjia Xu's page!

You can also find my publications on Google Scholar.

Conference Papers

  1. Zhengjia Xu, Dingyang Lyu, Zitai Qiu, Shan Xue, Jian Yang, and Jia Wu. “Is the Last Check-In All You Need? Next POI Recommendation: Recall and Rerank.” In: The 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD), 2026. (CCF-A, CORE-A*) [GitHub] [PDF]
  2. Jinghui Zhang, Zhengjia Xu, Dingyang Lv, Zhan Shi, Dian Shen, Jiahui Jin, and Fang Dong. “DiG-In-GNN: Discriminative Feature Guided GNN-Based Fraud Detector against Inconsistencies in Multi-Relation Fraud Graph.” In: The 38th AAAI Conference on Artificial Intelligence (AAAI), 2024. (CCF-A, CORE-A*) [GitHub]
  3. Caishan Weng, Zhiyang Shu, Zhengjia Xu, Jinghui Zhang, Junzhou Luo, Fang Dong, Peng Wang, and Zhengang Wang. “FreezePipe: An Efficient Dynamic Pipeline Parallel Approach Based on Freezing Mechanism for Distributed DNN Training.” In: The 26th International Conference on Computer Supported Cooperative Work in Design (CSCWD), 2023. (CCF-C, CORE-C)

Journal Papers

  1. Dingyang Lv, Zhengjia Xu, Jinghui Zhang, Yuchen Wang, and Fang Dong. “Imbalanced Node Classification with Graph Neural Networks: A Unified Approach Leveraging Homophily and Label Information.” In: Applied Soft Computing Journal, 2023. (JCR-Q1)
  2. Jinghui Zhang, Zhengjia Xu, Junhe Li, Qiangsheng Dai, Zhen Ling, and Ming Yang. “Wi-Fi device identification based on multi-domain physical layer fingerprint.” In: Computer Communications, 204, 118–129, 2023. (CCF-C, JCR-Q2)

Arxiv Papers

  1. Zhengjia Xu, Dingyang Lyu, and Jinghui Zhang. “Slicing Input Features to Accelerate Deep Learning: A Case Study with Graph Neural Networks.” In: arXiv:2408.11500 [Machine Learning (cs.LG)], 2024. [PDF]

Patents

  1. “A load-aware memory reallocation optimization method and system for heterogeneous acceleration device.” Jinghui Zhang, Caishan Weng, Zhengjia Xu, Jiange Li, Fang Dong, and Junzhou Luo. (CN115437795B)
  2. “A Wi-Fi device identification system and method based on multi-domain physical layer fingerprint features.” Jinghui Zhang, Caishan Weng, Zhengjia Xu, Zhen Ling, Ming Yang, and Junzhou Luo. (CN115664905A)