Graph Transfer Learning via Adversarial Domain Adaptation with Graph Convolution

2019/09/04 by Quanyu Dai, Xiao-Ming Wu, Dai, Quanyu +7 · 11 citations
Computer Science · Mathematics · #Adversarial system #Artificial intelligence #Artificial neural network #Computer science #Convolution (computer science) #Data mining #Discriminative model #Domain Adaptation and Few-Shot Learning #Domain adaptation #FOS: Computer and information sciences #Graph #Leverage (statistics) #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine learning #Negative transfer #Node (physics) #Social and Information Networks (cs.SI) #Source code #Theoretical computer science #Transfer of learning #cs.LG #cs.SI #stat.ML

paper · pdf · doi:10.48550/arxiv.1909.01541

published in arXiv (Cornell University) (Cornell University) · Accepted by IEEE Transactions on Knowledge and Data Engineering

openalex publication_date 2019/09/04 · arxiv created 2022/07/31 · arxiv updated 2022/08/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08

Abstract

This paper studies the problem of cross-network node classification to overcome the insufficiency of labeled data in a single network. It aims to leverage the label information in a partially labeled source network to assist node classification in a completely unlabeled or partially labeled target network. Existing methods for single network learning cannot solve this problem due to the domain shift across networks. Some multi-network learning methods heavily rely on the existence of cross-network connections, thus are inapplicable for this problem. To tackle this problem, we propose a novel \textcolorblackgraph transfer learning framework AdaGCN by leveraging the techniques of adversarial domain adaptation and graph convolution. It consists of two components: a semi-supervised learning component and an adversarial domain adaptation component. The former aims to learn class discriminative node representations with given label information of the source and target networks, while the latter contributes to mitigating the distribution divergence between the source and target domains to facilitate knowledge transfer. Extensive empirical evaluations on real-world datasets show that AdaGCN can successfully transfer class information with a low label rate on the source network and a substantial divergence between the source and target domains. The source code for reproducing the experimental results is available at https://github.com/daiquanyu/AdaGCN.

Cited by

• Empowering GNNs for Domain Adaptation via Denoising Target Graph
• Structure-Attribute Transformations with Markov Chain Boost Graph Domain Adaptation
• Sparse Causal Discovery with Generative Intervention for Unsupervised Graph Domain Adaptation
• Cross-Resolution Semantic Learning for Graph Domain Adaptation
• BotTrans: A Multi-Source Graph Domain Adaptation Approach for Social Bot Detection
• Bridging Source and Target Domains via Link Prediction for Unsupervised Domain Adaptation on Graphs
• Homophily Enhanced Graph Domain Adaptation
• Dynamic Bundling with Large Language Models for Zero-Shot Inference on Text-Attributed Graphs
• GCAL: Adapting Graph Models to Evolving Domain Shifts
• Graph Data Selection for Domain Adaptation: A Model-Free Approach
• Pave Your Own Path: Graph Gradual Domain Adaptation on Fused Gromov-Wasserstein Geodesics

Related

• Non-IID Transfer Learning on Graphs
• Semi-supervised Domain Adaptation in Graph Transfer Learning
• DANE: Domain Adaptive Network Embedding
• Transfer Learning Under High-Dimensional Graph Convolutional Regression Model for Node Classification
• Investigating Transfer Learning in Graph Neural Networks