用结构惩罚增强基于距离的图自编码器以进行动态图嵌入

View PDF HTML (experimental)

Abstract:Graph autoencoders (GAEs) are widely used for learning representations of dynamic graphs. However, their optimisation objectives typically do not take structural heterogeneity across nodes into account. We propose three distance-based GAE variants that incorporate structural penalties into the reconstruction loss. All variants share a two-layer Graph Convolutional Network encoder and a Euclidean-distance decoder trained with distance-based reconstruction objectives. We extend sparsity-corrected loss with two node-level regularization terms: (i) a hub penalty based on degree centrality, and (ii) a penalty based on Natural Community Local Intrinsic Dimensionality (NC-LID). The paper is motivated by prior evidence linking high NC-LID to reduced embedding quality. The proposed methods are designed to emphasize reconstruction errors for structurally ambiguous nodes. Experiments on multiple dynamic graph data sets show that incorporating NC-LID-based regularization consistently improves reconstruction performance over the baseline without structural regularization and the method using hub-aware regularization. These findings highlight NC-LID as a useful structural signal for enhancing distance-based graph autoencoders in dynamic settings.

Subjects:	Machine Learning (cs.LG); Emerging Technologies (cs.ET)
Cite as:	arXiv:2608.18762 [cs.LG]
	(or arXiv:2608.18762v2 [cs.LG] for this version)
	https://doi.org/10.48550/arXiv.2608.18762 arXiv-issued DOI via DataCite

Submission history

From: Aleksandar Tomčić [view email]
[v1] Wed, 19 Aug 2026 10:13:56 UTC (29 KB)
[v2] Tue, 6 Oct 2026 08:14:37 UTC (30 KB)