arXiv：cs.LG· Nguyen Duy Long, Phung Minh Hien, Nguyen Trong Viet, Nguyen Thai Anh·· 4 小时前AI 评分34
基于 TabICL 的免训练图节点分类：同质性门控的保形预测可靠性研究
Valid for Free: Homophily-Gated Conformal Prediction for Training-Free Node Classification with Tabular Foundation Models
AI 导读
研究首次对表格基础模型（TFM）免训练图节点分类进行保形可靠性分析，以 TabICL 为模型、每个图一半节点作标注上下文。在十个图上的审计显示，TabICL 后验的平均 ECE 为 0.019，比带温度缩放的 GCN（0.029）低约 35%。作者提出免训练扩散分数 HG-DAPS，在六个同质性图上将平均预测集大小比 APS 缩小 5.8% 至 17.1%。
正文
Abstract:Tabular foundation models (TFMs) can classify the nodes of a graph without training on it, by reading node and neighborhood features as table rows next to labeled context rows. Work in this line reports predictive performance, not conformal coverage or prediction-set size. To our knowledge, we give the first reliability study of the setting, with TabICL as the TFM and half of each graph as labeled context. As for any predictor fixed before calibration, a frozen in-context predictor makes split conformal prediction exactly valid in finite samples, with no training, validation fold, or tuning on the target graph. An audit across ten graphs then shows that the training-free TabICL posterior has lower expected calibration error (ECE) than GCN with temperature scaling (GCN+TS) on nine of them. Its mean ECE over the ten graphs is 0.019, about 35 percent below the 0.029 of GCN+TS. We also introduce HG-DAPS, a training-free diffusion score whose homophily gate reads only the in-context labels, so the guarantee still holds. Relative to adaptive prediction sets (APS), it reduces mean set size by 5.8 to 17.1 percent on six homophilous graphs and changes it by under 1 percent on four heterophilous ones. On two binary, class-imbalanced graphs, a pre-registered trap case shows that gating on raw rather than adjusted homophily lowers coverage among low-homophily nodes by 0.27 and 0.12. Marginal coverage stays at the nominal 0.90 and masks this drop.
| Comments: | 6 pages, 4 figures, 1 table |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.08564 [cs.LG] |
| (or arXiv:2610.08564v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.08564 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Nguyen Thai Anh [view email]
[v1]
Tue, 6 Oct 2026 15:44:07 UTC (279 KB)
来源：arXiv：cs.LG · arxiv.org