Saved in:
Bibliographic Details
Main Authors: Luo, Junyu, Tang, Yuhao, Fu, Yiwei, Luo, Xiao, Kou, Zhizhuo, Xiao, Zhiping, Ju, Wei, Zhang, Wentao, Zhang, Ming
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.07621
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908444389277696
author Luo, Junyu
Tang, Yuhao
Fu, Yiwei
Luo, Xiao
Kou, Zhizhuo
Xiao, Zhiping
Ju, Wei
Zhang, Wentao
Zhang, Ming
author_facet Luo, Junyu
Tang, Yuhao
Fu, Yiwei
Luo, Xiao
Kou, Zhizhuo
Xiao, Zhiping
Ju, Wei
Zhang, Wentao
Zhang, Ming
contents Unsupervised Graph Domain Adaptation (UGDA) leverages labeled source domain graphs to achieve effective performance in unlabeled target domains despite distribution shifts. However, existing methods often yield suboptimal results due to the entanglement of causal-spurious features and the failure of global alignment strategies. We propose SLOGAN (Sparse Causal Discovery with Generative Intervention), a novel approach that achieves stable graph representation transfer through sparse causal modeling and dynamic intervention mechanisms. Specifically, SLOGAN first constructs a sparse causal graph structure, leveraging mutual information bottleneck constraints to disentangle sparse, stable causal features while compressing domain-dependent spurious correlations through variational inference. To address residual spurious correlations, we innovatively design a generative intervention mechanism that breaks local spurious couplings through cross-domain feature recombination while maintaining causal feature semantic consistency via covariance constraints. Furthermore, to mitigate error accumulation in target domain pseudo-labels, we introduce a category-adaptive dynamic calibration strategy, ensuring stable discriminative learning. Extensive experiments on multiple real-world datasets demonstrate that SLOGAN significantly outperforms existing baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2507_07621
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Sparse Causal Discovery with Generative Intervention for Unsupervised Graph Domain Adaptation
Luo, Junyu
Tang, Yuhao
Fu, Yiwei
Luo, Xiao
Kou, Zhizhuo
Xiao, Zhiping
Ju, Wei
Zhang, Wentao
Zhang, Ming
Machine Learning
Unsupervised Graph Domain Adaptation (UGDA) leverages labeled source domain graphs to achieve effective performance in unlabeled target domains despite distribution shifts. However, existing methods often yield suboptimal results due to the entanglement of causal-spurious features and the failure of global alignment strategies. We propose SLOGAN (Sparse Causal Discovery with Generative Intervention), a novel approach that achieves stable graph representation transfer through sparse causal modeling and dynamic intervention mechanisms. Specifically, SLOGAN first constructs a sparse causal graph structure, leveraging mutual information bottleneck constraints to disentangle sparse, stable causal features while compressing domain-dependent spurious correlations through variational inference. To address residual spurious correlations, we innovatively design a generative intervention mechanism that breaks local spurious couplings through cross-domain feature recombination while maintaining causal feature semantic consistency via covariance constraints. Furthermore, to mitigate error accumulation in target domain pseudo-labels, we introduce a category-adaptive dynamic calibration strategy, ensuring stable discriminative learning. Extensive experiments on multiple real-world datasets demonstrate that SLOGAN significantly outperforms existing baselines.
title Sparse Causal Discovery with Generative Intervention for Unsupervised Graph Domain Adaptation
topic Machine Learning
url https://arxiv.org/abs/2507.07621