发表机构
State Key Laboratory of Complex & Critical Software Environment, Beihang University; Shen Yuan Honors College, Beihang University（北京航空航天大学复杂关键软件环境国家重点实验室; 北京航空航天大学沈元荣誉学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出免训练的EMC框架，通过闭式解将图域结晶为紧凑记忆，以高效应对非平稳分布偏移，显著降低运行时间和内存消耗。
AI 中文摘要
部署在真实系统中的深度图学习模型通常需要应对非平稳环境，其中底层图分布随时间持续漂移。现有解决方案依赖于训练辅助生成模块来合成记忆图以进行跨域适应，这会产生大量计算开销，并且在长期分布偏移下扩展性较差。我们认为存在一条更经济的路径：与其生成记忆，不如将其结晶化。为此，我们提出了高效记忆结晶化（EMC），这是一种免训练、测试时框架，通过记忆导向的分布匹配目标的闭式解，将每个到来的图域提炼为紧凑、语义忠实的记忆，从而在持续协变量偏移下消除冗余的域信息。为了在模型遍历长序列的目标域时保持泛化性和适应性，EMC进一步通过状态演化的记忆建模域间依赖，并获得了理论上更紧的泛化误差界，优于直接适应。大量实验表明，在非平稳分布偏移下的图上，EMC相对于最先进的基线表现出优越性能，同时与最近的竞争对手相比，平均运行时间减少了87.4%，GPU内存消耗减少了92.4%，使得持续图适应在大规模上变得实用。
英文摘要
Deep graph learning models deployed in real-world systems often need to cope with non-stationary environments, where the underlying graph distribution drifts continually over time. Prevailing solutions rely on training auxiliary generative modules to synthesize memory graphs for cross-domain adaptation, which incurs substantial computational overhead and scales poorly under prolonged distribution shifts. We argue that a more economical path exists: rather than generating memory, one can crystallize it. To this end, we propose Efficient Memory Crystallization (EMC), a training-free test-time framework that distills each incoming graph domain into a compact, semantically faithful memory through a closed-form solution to a memory-oriented distribution-matching objective, thereby eliminating redundant domain information under continual covariate shifts. To preserve both generalizability and adaptability as the model traverses a long sequence of target domains, EMC further models inter-domain dependencies through state-evolving memories and admits a theoretically grounded, tighter generalization error bound than direct adaptation. Extensive experiments demonstrate the superior performance of EMC over state-of-the-art baselines on graphs under non-stationary distribution shifts, while reducing average runtime by 87.4% and GPU memory consumption by 92.4% relative to the recent competitor, making continual graph adaptation practical at scale.
CommentsAccepted by the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)
