arXiv cs.AI· Joo Young Park, Sejun Park, Hyunwoo Park·· 2 天前AI 评分22
STAMP：面向时间切分图域适应的对称时间平均消息传递
STAMP: Symmetric Temporal Averaging Message Passing for Domain Adaptation in Chronologically Split Graphs
AI 导读
针对时间切分图下的域适应问题，研究者提出对称时间平均消息传递算子 STAMP，用跨时间戳平衡的消息传播替代标准邻域聚合，强制时间域间一阶矩不变。其异构变体 STAMP-mag 在 ogbn-mag 上取得 SOTA，STAMP 在 ogbn-arxiv 上取得 SOTA，并在时间切分 Pokec 上持续提升原数据集 SOTA 模型的表现。
正文 · 原文
Abstract:This paper addresses domain adaptation in graph-structured data under chronological splits. We consider transductive semi-supervised node classification (SSNC), in which labels of historical nodes are used to predict labels of more recent nodes. Temporal dependencies in graph connectivity induce distribution shifts across time, often degrading the performance of models trained on past nodes. Despite its practical relevance, domain adaptation under chronological graph splits remains relatively underexplored. We propose Symmetric Temporal Averaging Message Passing (STAMP), a timestamp-balanced operator that replaces standard neighborhood aggregation with temporally aligned propagation. We call its specialized variant for the heterogeneous ogbn-mag dataset STAMP-mag, reserving STAMP for the general formulation. By balancing messages across timestamps, STAMP enforces first-moment invariance across temporal domains while remaining scalable and easy to implement. Unlike many existing domain adaptation methods, whose assumptions are difficult to verify on graphs, STAMP relies on transparent, practically checkable conditions and is supported by theoretical analysis. Empirically, STAMP and STAMP-mag improve accuracy across all evaluated temporal graph benchmarks. STAMP achieves state-of-the-art performance on ogbn-arxiv, while STAMP-mag does so on ogbn-mag. For temporal-split Pokec, where no established benchmark exists, applying STAMP to the state-of-the-art model for the original dataset consistently improves performance under our chronological split. These results demonstrate that timestamp-balanced propagation is an effective and practical approach to mitigating temporal distribution shift in graph learning.
| Comments: | 25 pages (without appendix), 34 pages (with appendix), 9 figures |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Machine Learning (stat.ML) |
| Cite as: | arXiv:2411.10957 [cs.LG] |
| (or arXiv:2411.10957v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2411.10957 arXiv-issued DOI via DataCite |
Submission history
From: Joo Young Park [view email]
[v1]
Sun, 17 Nov 2024 04:23:25 UTC (1,941 KB)
[v2]
Wed, 14 Oct 2026 03:57:22 UTC (1,815 KB)
来源：arXiv cs.AI · arxiv.org