GraphKeeper: Graph Domain-Incremental Learning via Knowledge Disentanglement and Preservation

2025-10-30

AI Summary

This paper presents the first systematic study of catastrophic forgetting in graph-domain incremental learning (Domain-IL). Addressing the limitations of existing methods—which are confined to single-domain task/class incremental settings and ill-suited for continual updates across heterogeneous graph domains—we propose a knowledge disentanglement and unbiased preservation framework. Our approach features: (i) a domain-aware, parameter-efficient fine-tuning mechanism that jointly optimizes intra-domain discriminability and inter-domain invariance; (ii) disentangled embedding-space constraints and bias-free knowledge distillation to mitigate embedding shift and decision boundary drift; and (iii) support for unknown domain detection. Extensive experiments on multiple graph benchmarks demonstrate substantial improvements over state-of-the-art methods, with average accuracy gains of 6.5%–16.6%, negligible forgetting rates, and seamless plug-and-play compatibility with mainstream graph foundation models.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Graph incremental learning (GIL), which continuously updates graph models by sequential knowledge acquisition, has garnered significant interest recently. However, existing GIL approaches focus on task-incremental and class-incremental scenarios within a single domain. Graph domain-incremental learning (Domain-IL), aiming at updating models across multiple graph domains, has become critical with the development of graph foundation models (GFMs), but remains unexplored in the literature. In this paper, we propose Graph Domain-Incremental Learning via Knowledge Dientanglement and Preservation (GraphKeeper), to address catastrophic forgetting in Domain-IL scenario from the perspectives of embedding shifts and decision boundary deviations. Specifically, to prevent embedding shifts and confusion across incremental graph domains, we first propose the domain-specific parameter-efficient fine-tuning together with intra- and inter-domain disentanglement objectives. Consequently, to maintain a stable decision boundary, we introduce deviation-free knowledge preservation to continuously fit incremental domains. Additionally, for graphs with unobservable domains, we perform domain-aware distribution discrimination to obtain precise embeddings. Extensive experiments demonstrate the proposed GraphKeeper achieves state-of-the-art results with 6.5%~16.6% improvement over the runner-up with negligible forgetting. Moreover, we show GraphKeeper can be seamlessly integrated with various representative GFMs, highlighting its broad applicative potential.

Problem

• Addressing catastrophic forgetting in graph domain-incremental learning scenarios
• Preventing embedding shifts and decision boundary deviations across domains
• Enabling continuous model updates across multiple graph domains

Innovation

• Domain-specific fine-tuning with disentanglement objectives
• Deviation-free knowledge preservation for stable boundaries
• Domain-aware distribution discrimination for unobservable graphs