arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2602.10489cs.LGcs.AI
收藏

基于神经特征函数的自适应分布对齐用于图域适应

Learning Adaptive Distribution Alignment with Neural Characteristic Function for Graph Domain Adaptation

School of Artificial Intelligence, Beihang University(北航人工智能学院);School of Computer Science and Engineering, Beihang University(北航计算机科学与工程学院);School of Mathematical Sciences, Peking University(北大数学系);Zhongguancun Laboratory(中关村实验室)

Wei Chen, Xingyu Guo, Shuang Li, Zhao Zhang, Yan Zhong, Fuzhen Zhuang, Deqing wang

AI总结:

本文提出ADAlign框架,通过自适应分布对齐解决图域适应中的分布偏移问题,利用神经谱差异衡量跨图偏移,实现灵活且鲁棒的域适应。

摘要:

图域适应(GDA)将知识从标记源图转移到未标记目标图,但面临复杂的多维分布偏移挑战。现有方法通过对齐手动选择的图元素(如节点属性或结构统计)来减少偏移,但需要手动设计的图滤波器提取相关特征。然而,这些方法缺乏灵活性:依赖场景特定的启发式方法,在不同转移场景中表现不佳。为了解决这些限制,我们提出ADAlign,一种适用于GDA的自适应分布对齐框架。与启发式方法不同,ADAlign无需手动指定对齐标准。它自动识别每个转移中最相关的偏移,并联合对齐它们,捕捉属性、结构及其依赖之间的相互作用。这使ADAlign灵活、场景感知且对多样且动态变化的偏移具有鲁棒性。为了实现这种适应性,我们引入了神经谱差异(NSD),一种理论上严谨的参数化距离,提供跨图偏移的统一视角。NSD利用谱域中的神经特征函数编码所有阶次的特征-结构依赖,同时可学习的频率采样器通过极小极大范式自适应强调每个任务中最信息丰富的频谱成分。在10个数据集和16个转移任务上的广泛实验表明,ADAlign不仅优于最先进的基线,还通过更低的内存使用和更快的训练实现了效率提升。

英文摘要:

Graph Domain Adaptation (GDA) transfers knowledge from labeled source graphs to unlabeled target graphs but is challenged by complex, multi-faceted distributional shifts. Existing methods attempt to reduce distributional shifts by aligning manually selected graph elements (e.g., node attributes or structural statistics), which typically require manually designed graph filters to extract relevant features before alignment. However, such approaches are inflexible: they rely on scenario-specific heuristics, and struggle when dominant discrepancies vary across transfer scenarios. To address these limitations, we propose \textbf{ADAlign}, an Adaptive Distribution Alignment framework for GDA. Unlike heuristic methods, ADAlign requires no manual specification of alignment criteria. It automatically identifies the most relevant discrepancies in each transfer and aligns them jointly, capturing the interplay between attributes, structures, and their dependencies. This makes ADAlign flexible, scenario-aware, and robust to diverse and dynamically evolving shifts. To enable this adaptivity, we introduce the Neural Spectral Discrepancy (NSD), a theoretically principled parametric distance that provides a unified view of cross-graph shifts. NSD leverages neural characteristic function in the spectral domain to encode feature-structure dependencies of all orders, while a learnable frequency sampler adaptively emphasizes the most informative spectral components for each task via minimax paradigm. Extensive experiments on 10 datasets and 16 transfer tasks show that ADAlign not only outperforms state-of-the-art baselines but also achieves efficiency gains with lower memory usage and faster training.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑