arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2602.09065cs.LGcs.AI
收藏

增强型图变换器与序列化图标记

Enhanced Graph Transformer with Serialized Graph Tokens

MAIS, Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所MAIS)

Ruixiang Wang, Yuyang Hong, Shiming Xiang, Chunhong Pan

AI总结:

本文提出一种增强型图变换器,通过序列化图标记方法提升图层面表示能力,在多个图学习基准上取得最优性能。

摘要:

变换器在图学习中已展现出成功,特别是在节点层面的任务中。然而,现有方法在生成图层面表示时遇到信息瓶颈。普遍的单标记范式无法充分利用自注意力机制在编码标记序列中的固有优势,并退化为节点信号的加权和。为了解决这个问题,我们设计了一种新颖的序列化标记范式,以更有效地封装全局信号。具体来说,提出了一种图序列化方法,将节点信号聚合为序列化图标记,其中位置编码被自动包含。然后,堆叠的自注意力层被应用于编码此标记序列并捕捉其内部依赖性。我们的方法可以通过建模多个图标记之间的复杂相互作用来生成更具表现力的图表示。实验结果表明,我们的方法在几个图层面基准上实现了最先进的结果。消融研究验证了所提出模块的有效性。

英文摘要:

Transformers have demonstrated success in graph learning, particularly for node-level tasks. However, existing methods encounter an information bottleneck when generating graph-level representations. The prevalent single token paradigm fails to fully leverage the inherent strength of self-attention in encoding token sequences, and degenerates into a weighted sum of node signals. To address this issue, we design a novel serialized token paradigm to encapsulate global signals more effectively. Specifically, a graph serialization method is proposed to aggregate node signals into serialized graph tokens, with positional encoding being automatically involved. Then, stacked self-attention layers are applied to encode this token sequence and capture its internal dependencies. Our method can yield more expressive graph representations by modeling complex interactions among multiple graph tokens. Experimental results show that our method achieves state-of-the-art results on several graph-level benchmarks. Ablation studies verify the effectiveness of the proposed modules.

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。

↑