arXiv：cs.AI· Firas Bayram, Maxime Cordy·· 6 小时前
CB-RBM：无需标签的属性图对抗检测与漂移稳定性
Feature-Context Consistency: Unsupervised Adversarial Detection with Drift Stability on Attributed Graphs
AI 导读
研究者提出 CB-RBM，将节点属性与邻域聚合属性之间的一致性形式化为 Context Boost（CB）分数，在 Restricted Boltzmann Machine 框架下实现无标签对抗检测并给出最坏情况保证。
正文
Abstract:Adversarial detection on evolving attributed graphs faces two challenges: fraud labels arrive late or not at all, and natural distributional drift erodes any trained decision boundary. Modern detectors can reach strong detection rates when labels are available and the distribution is stationary; sustaining that quality without labels and under drift is the open problem. We identify feature-context consistency -- the alignment between a node's attributes and its aggregated neighborhood attributes -- as a single structural signal that addresses both. We formalize it as the Context Boost (CB) score within a Restricted Boltzmann Machine framework (CB-RBM) and prove two complementary worst-case guarantees: a strict separation between the expected CB of legitimate and adversarial nodes, with gap growing linearly in model capacity, whenever the adversary's context stays far enough from legitimate neighborhood context; and a Wasserstein-Lipschitz bound on how far the CB distribution moves under natural drift. We validate on XBlock Ethereum phishing (~153K nodes) and Reddit banned-user detection (~11K nodes), with two further datasets in the appendix spanning diverse domains. With zero labels, CB-RBM achieves AUROC >= 0.96 on four adversarial attacks and keeps FPR close to its calibrated 5% target under drift -- outperforming supervised GNNs (hundreds of labels), unsupervised graph-anomaly baselines, and a standard RBM whose FPR collapses to 89-100%. What typically requires separate mechanisms -- label-free operation and drift stability -- here follows from a single structural regularity of the data. Code: this https URL.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.16694 [cs.LG] |
| (or arXiv:2610.16694v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.16694 arXiv-issued DOI via DataCite (pending registration) |
|
| Journal reference: | Advances in Neural Information Processing Systems 39 (NeurIPS 2026) |
Submission history
From: Firas Bayram [view email]
[v1]
Mon, 12 Oct 2026 12:30:20 UTC (1,729 KB)
来源：arXiv：cs.AI · arxiv.org