arXiv：cs.LG· Louis Tichelman, Xingyue Huang, Jinwoo Kim, \.Ismail \.Ilkan Ceylan·· 6 小时前AI 评分42
Wander：用随机游走跨图与任务学习的图基础模型
To Learn is to Wander: Learning Across Graphs and Tasks with Random Walks
AI 导读
研究者提出图基础模型 Wander，将图学习建模为部分观测图的补全，通过随机游走这一通用接口，让单一预训练 checkpoint 同时适配同构图与多关系图。Wander 可在推理时增加结构上下文而不改变已学参数，并在有界连通图上通用逼近相应的贝叶斯最优预测器。
正文
Abstract:Graph foundation models aim to transfer across graphs, feature spaces, relational schemas, and prediction tasks, yet existing approaches typically generalize only within particular graph modalities or tasks. We propose Wander, a graph foundation model designed to operate across these settings within a single pretrained checkpoint. Following the prior-predictive perspective, we formulate graph learning as completion of a partially observed graph. We realize this task-general view through a common interface based on random walks, allowing the same model to operate across homogeneous and multi-relational graphs with varying features, labels, and relational schemas. Wander can increase its structural context at inference time without changing its learned parameters and, under suitable assumptions, universally approximates the corresponding Bayes-optimal predictor on bounded connected graphs. Empirically, a single pretrained checkpoint achieves state-of-the-art or highly competitive results across node classification, homogeneous link prediction, and knowledge-graph link prediction. Moreover, joint pretraining across graph modalities and tasks preserves performance in specialized settings while enabling positive transfer and the composition of separately learned capabilities at inference time.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.06694 [cs.LG] |
| (or arXiv:2610.06694v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.06694 arXiv-issued DOI via DataCite |
Submission history
From: Louis Tichelman [view email]
[v1]
Mon, 5 Oct 2026 16:55:05 UTC (6,444 KB)
[v2]
Tue, 6 Oct 2026 11:52:34 UTC (6,444 KB)
来源：arXiv：cs.LG · arxiv.org