NeurIPS 2022 · Spotlight (5.8%)
Self-supervised Heterogeneous Graph Pre-training based on Structural Clustering
Yaming Yang, Ziyu Guan, Zhe Wang, Wei Zhao*, Cai Xu, Weigang Lu, Jianbin Huang
Proposed SHGP — a self-supervised pre-training approach that requires no positive/negative sample generation. Achieved 98.41% node classification accuracy on MAG dataset. Widely reported by AI media outlets.
WWW 2026 · CCF-A
Aligning Multiple Knowledge Graphs in A Single Pass
Yaming Yang, Zhe Wang, Ziyu Guan*, Wei Zhao, Weigang Lu, Xinyan Huang, Jiangtao Cui, Xiaofei He
First work to study and solve multi-KG (>2) alignment in a single pass. Proposed MultiEA framework with three alignment strategies and high-order similarity inference enhancement. Built two new benchmark datasets.
IEEE TKDE 2025 · CCF-A Journal
Unsupervised Entity Alignment Based on Personalized Discriminative Rooted Tree
Yaming Yang, Zhe Wang, Ziyu Guan, Wei Zhao*, Xinyan Huang, Xiaofei He
Proposed UNEA with parametric tree neighborhood sampling and tree attention aggregation for personalized embeddings. Outperformed many supervised EA baselines in unsupervised settings.
ICDE 2025 · CCF-A
A Translation-based Heterogeneous Graph Neural Network for Multiple Knowledge Graphs Alignment
Yaming Yang, Zhuofeng Luo, Zhe Wang, Weigang Lu, Yiheng Lu, Ziyu Guan*, Wei Zhao, Yuanhai Lv
Proposed KG2HIN that transforms KGs into HINs, combining HGNN aggregation with translation-based KG embedding. Improved M-Hits@1 from 10.25% to 73.05% on DBP-4 with fewer parameters.
NeurIPS 2025 · CCF-A
Defining and Discovering Hyper-meta-paths for Heterogeneous Hypergraphs
Yaming Yang, Ziyu Zheng, Weigang Lu, Zhe Wang, Xinyan Huang, Wei Zhao, Ziyu Guan*
Introduced hyper-meta-path concept for heterogeneous hypergraphs. Designed HHNN with attention to learn importance of hyper-meta-paths, improving both performance and interpretability.
CIKM 2025 · CCF-B
Audience-Aware and Self-Adaptive Multi-Interest Modeling for Sharing Rate Prediction in Affiliate Marketing
Zhe Wang, Yaming Yang, Ziyu Guan*, Rui Wang, Yujian Cao, Bin Tong*, Wei Zhao, Hongbo Deng*
Multi-interest modeling for B-end promoter sharing rate prediction. Proposed dynamic routing with interest capsules and dual-channel attention for audience preference integration.
CIKM 2025 · CCF-B
Dynamic Network-Based Two-Stage Time Series Forecasting for Affiliate Marketing
Zhe Wang, Yaming Yang, Bin Tong, Rui Wang, Ziyu Guan, Wei Zhao*, Hongbo Deng*
Designed propagation scale metric for promoter indirect contributions. Two-stage solution with descendant graph convolution encoding and hypergraph convolution.