Nested Graph Pseudo-Label Refinement for Noisy Label Domain Adaptation Learning

2025-08-01

AI Summary

In graph domain adaptation (GDA), prevalent label noise in the source domain severely degrades feature alignment and overall performance. Method: We propose a noise-robust dual-branch GDA framework comprising semantic and topological branches, incorporating neighborhood consistency constraints to mitigate noise interference. We design a nested pseudo-label refinement mechanism to enable progressive cross-domain learning and—novelly—integrate this nested optimization with theoretically grounded, noise-aware regularization to jointly suppress source-domain overfitting and pseudo-label noise in the target domain. Contribution/Results: Evaluated on multiple benchmark datasets, our method achieves up to 12.7% absolute accuracy improvement over state-of-the-art approaches, demonstrating superior robustness—particularly under high-noise conditions—while maintaining theoretical guarantees for noise resilience.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Adversarial Attacks & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Graph Domain Adaptation (GDA) facilitates knowledge transfer from labeled source graphs to unlabeled target graphs by learning domain-invariant representations, which is essential in applications such as molecular property prediction and social network analysis. However, most existing GDA methods rely on the assumption of clean source labels, which rarely holds in real-world scenarios where annotation noise is pervasive. This label noise severely impairs feature alignment and degrades adaptation performance under domain shifts. To address this challenge, we propose Nested Graph Pseudo-Label Refinement (NeGPR), a novel framework tailored for graph-level domain adaptation with noisy labels. NeGPR first pretrains dual branches, i.e., semantic and topology branches, by enforcing neighborhood consistency in the feature space, thereby reducing the influence of noisy supervision. To bridge domain gaps, NeGPR employs a nested refinement mechanism in which one branch selects high-confidence target samples to guide the adaptation of the other, enabling progressive cross-domain learning. Furthermore, since pseudo-labels may still contain noise and the pre-trained branches are already overfitted to the noisy labels in the source domain, NeGPR incorporates a noise-aware regularization strategy. This regularization is theoretically proven to mitigate the adverse effects of pseudo-label noise, even under the presence of source overfitting, thus enhancing the robustness of the adaptation process. Extensive experiments on benchmark datasets demonstrate that NeGPR consistently outperforms state-of-the-art methods under severe label noise, achieving gains of up to 12.7% in accuracy.

Problem

• Addresses noisy label issues in graph domain adaptation
• Proposes nested pseudo-label refinement for cross-domain learning
• Enhances robustness against source label noise and overfitting

Innovation

• Dual-branch pretraining with neighborhood consistency
• Nested refinement for cross-domain learning
• Noise-aware regularization against pseudo-label noise