GCAL: Adapting Graph Models to Evolving Domain Shifts
This addresses the challenge of adapting graph models to continuous domain shifts, which is crucial for applications like social networks or recommendation systems, though it is incremental as it builds on existing graph domain adaptation techniques.
The paper tackles the problem of graph domain adaptation for evolving out-of-distribution graphs, where conventional methods fail due to single-step adaptation and catastrophic forgetting. It introduces GCAL, which uses bilevel optimization with information maximization and variational memory generation, achieving substantial improvements in adaptability and knowledge retention over existing methods.
This paper addresses the challenge of graph domain adaptation on evolving, multiple out-of-distribution (OOD) graphs. Conventional graph domain adaptation methods are confined to single-step adaptation, making them ineffective in handling continuous domain shifts and prone to catastrophic forgetting. This paper introduces the Graph Continual Adaptive Learning (GCAL) method, designed to enhance model sustainability and adaptability across various graph domains. GCAL employs a bilevel optimization strategy. The "adapt" phase uses an information maximization approach to fine-tune the model with new graph domains while re-adapting past memories to mitigate forgetting. Concurrently, the "generate memory" phase, guided by a theoretical lower bound derived from information bottleneck theory, involves a variational memory graph generation module to condense original graphs into memories. Extensive experimental evaluations demonstrate that GCAL substantially outperforms existing methods in terms of adaptability and knowledge retention.