FedOGL: Combating Catastrophic Forgetting in Federated Open-World Multimodal Graph Learning

2026-07-30

AI Summary

This work addresses catastrophic forgetting in federated open-world multimodal graph learning, which arises from modality semantic coverage gaps, topological structure erosion, and fragmented client-side memory. To mitigate these issues, the authors propose FedOGL, a novel framework that integrates experience replay with task-initiated distillation at the client level to preserve historical decisions, while projecting graph propagation memory onto a globally shared structural basis. At the server level, FedOGL maintains and distributes compact class prototypes to enable privacy-preserving knowledge sharing across clients. This is the first approach to jointly tackle both structural and semantic forgetting through structural basis projection and prototype sharing mechanisms. Experimental results demonstrate that FedOGL reduces performance degradation due to forgetting by 42.67% compared to the strongest baseline, while maintaining or improving overall performance on downstream tasks.

Abstract

Federated graph learning enables collaborative training over decentralized graph data without sharing raw graph information. As such risks evolve, clients must learn emerging classes from private multimodal graph streams, retain historical categories, and reject samples outside the known class space. In this setting, clients must learn emerging classes from private multimodal graph streams while preserving historical categories and rejecting samples outside the current known class space. The core challenge is catastrophic forgetting, which in federated multimodal graphs is not merely a classifier-level failure: old knowledge can be erased through modality-semantic overwriting, topology-induced structural erosion, and federated memory fragmentation. To address this challenge, we propose \textbf{FedOGL}, a semantic-structural memory preservation framework. On the client side, FedOGL preserves historical decision behavior through replay and task-start distillation, while protecting graph-propagation memory via projection onto a globally shared structure basis. On the server side, FedOGL maintains and transfers compact category prototypes to facilitate cross-client knowledge sharing without exposing raw graph data. Extensive experiments demonstrate that, compared with the best-performing baselines, FedOGL reduces performance degradation caused by catastrophic forgetting by \textbf{42.67\%}, while maintaining or improving performance on downstream tasks.

Problem

• catastrophic forgetting
• federated learning
• open-world learning
• multimodal graph learning
• class-incremental learning

Innovation

• Federated Learning
• Catastrophic Forgetting
• Multimodal Graph Learning
• Memory Preservation
• Open-World Learning