Generalized Category Discovery in Federated Graph Learning

2026-05-05

AI Summary

Existing federated graph learning methods are constrained by the closed-world assumption, struggling to accommodate dynamically emerging classes and suffering from neighborhood absorption effects and cross-client semantic inconsistency. This work introduces generalized category discovery into federated graph learning for the first time, proposing a joint optimization framework: clients mitigate neighborhood bias through topology-aware semantic alignment, while the server addresses semantic drift under heterogeneous subgraphs via a hierarchical prototype alignment strategy. Evaluated on five real-world graph datasets, the proposed method significantly outperforms current approaches, achieving an average absolute improvement of 4.86 in HRScore and effectively enabling collaborative recognition of both known and novel categories.

Abstract

Federated Graph Learning (FGL) enables collaborative learning over distributed graph data, yet existing approaches largely rely on a closed-world assumption, limiting their applicability in dynamic environments where novel categories continuously emerge. To bridge this gap, we target the practical scenario of Federated Graph Generalized Category Discovery (FGGCD), aiming to collaboratively discover novel categories across decentralized graph clients while retaining knowledge of known categories. We observe that FGGCD introduces two fundamental challenges: (1) the Neighborhood Absorption Effect, where structural fragmentation leads to biased neighborhood aggregation, causing novel nodes to be misclassified as known categories; and (2) Global Semantic Inconsistency, where the aforementioned local biases propagate to the server and are amplified by heterogeneous subgraph distributions, hindering cross-client knowledge integration. To address these issues, we propose GCD-FGL, an FGL framework for GCD that integrates a client-side Topology-Reliable Semantic Alignment and Discovery process to mitigate the neighborhood absorption effect, and a server-side Hierarchical Prototype Alignment strategy to resolve global semantic inconsistency. Extensive experiments on five real-world graph datasets demonstrate that GCD-FGL consistently outperforms state-of-the-art baselines, achieving an average absolute gain of +4.86 in HRScore.

Problem

• Federated Graph Learning
• Generalized Category Discovery
• Novel Category Discovery
• Decentralized Graph Data
• Dynamic Environments

Innovation

• Federated Graph Learning
• Generalized Category Discovery
• Neighborhood Absorption Effect
• Hierarchical Prototype Alignment
• Semantic Consistency