On the Limitation and Experience Replay for GNNs in Continual Learning

TL;DR

This paper provides the first theoretical analysis of GNN learnability in continual learning, highlighting the impact of structural shifts, and introduces SEA-ER, a novel experience replay method that mitigates structural shift effects.

Contribution

It offers the first theoretical exploration of GNN learnability in NGCL and proposes SEA-ER, a new experience replay strategy that addresses structural shifts in graph data.

Findings

GNN learnability is heavily affected by structural shifts.

SEA-ER effectively reduces catastrophic forgetting in NGCL.

Experimental results validate the theoretical analysis and the effectiveness of SEA-ER.

Abstract

Continual learning seeks to empower models to progressively acquire information from a sequence of tasks. This approach is crucial for many real-world systems, which are dynamic and evolve over time. Recent research has witnessed a surge in the exploration of Graph Neural Networks (GNN) in Node-wise Graph Continual Learning (NGCL), a practical yet challenging paradigm involving the continual training of a GNN on node-related tasks. Despite recent advancements in continual learning strategies for GNNs in NGCL, a thorough theoretical understanding, especially regarding its learnability, is lacking. Learnability concerns the existence of a learning algorithm that can produce a good candidate model from the hypothesis/weight space, which is crucial for model selection in NGCL development. This paper introduces the first theoretical exploration of the learnability of GNN in NGCL, revealing that learnability is heavily influenced by structural shifts due to the interconnected nature of graph data. Specifically, GNNs may not be viable for NGCL under significant structural changes, emphasizing the need to manage structural shifts. To mitigate the impact of structural shifts, we propose a novel experience replay method termed Structure-Evolution-Aware Experience Replay (SEA-ER). SEA-ER features an innovative experience selection strategy that capitalizes on the topological awareness of GNNs, alongside a unique replay strategy that employs structural alignment to effectively counter catastrophic forgetting and diminish the impact of structural shifts on GNNs in NGCL. Our extensive experiments validate our theoretical insights and the effectiveness of SEA-ER.