GIDN: A Lightweight Graph Inception Diffusion Network for High-efficient Link Prediction
Zixiao Wang, Yuluo Guo, Jin Zhao, Yu Zhang, Hui Yu, Xiaofei Liao, Biao Wang, Ting Yu
2022-10-04Link Prediction
Abstract
In this paper, we propose a Graph Inception Diffusion Networks(GIDN) model. This model generalizes graph diffusion in different feature spaces, and uses the inception module to avoid the large amount of computations caused by complex network structures. We evaluate GIDN model on Open Graph Benchmark(OGB) datasets, reached an 11% higher performance than AGDN on ogbl-collab dataset.
Results
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Link Property Prediction | ogbl-ddi | Number of params | 3506691 | GIDN@YITU |
| Link Property Prediction | ogbl-collab | Number of params | 60449025 | GIDN@YITU |
Related Papers
Topic Modeling and Link-Prediction for Material Property Discovery2025-07-08Graph Collaborative Attention Network for Link Prediction in Knowledge Graphs2025-07-05Understanding Generalization in Node and Link Prediction2025-07-01Context-Driven Knowledge Graph Completion with Semantic-Aware Relational Message Passing2025-06-29Directed Link Prediction using GNN with Local and Global Feature Fusion2025-06-25Demystifying Distributed Training of Graph Neural Networks for Link Prediction2025-06-25Call Me Maybe: Enhancing JavaScript Call Graph Construction using Graph Neural Networks2025-06-22A Transfer Learning Framework for Multilayer Networks via Model Averaging2025-06-14