Saved in:
Bibliographic Details
Main Authors: Ceccherini, Emma, Gallagher, Ian, Jones, Andrew, Lawson, Daniel
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.02859
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912459129880576
author Ceccherini, Emma
Gallagher, Ian
Jones, Andrew
Lawson, Daniel
author_facet Ceccherini, Emma
Gallagher, Ian
Jones, Andrew
Lawson, Daniel
contents Stability for dynamic network embeddings ensures that nodes behaving the same at different times receive the same embedding, allowing comparison of nodes in the network across time. We present attributed unfolded adjacency spectral embedding (AUASE), a stable unsupervised representation learning framework for dynamic networks in which nodes are attributed with time-varying covariate information. To establish stability, we prove uniform convergence to an associated latent position model. We quantify the benefits of our dynamic embedding by comparing with state-of-the-art network representation learning methods on four real attributed networks. To the best of our knowledge, AUASE is the only attributed dynamic embedding that satisfies stability guarantees without the need for ground truth labels, which we demonstrate provides significant improvements for link prediction and node classification.
format Preprint
id arxiv_https___arxiv_org_abs_2503_02859
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Unsupervised Attributed Dynamic Network Embedding with Stability Guarantees
Ceccherini, Emma
Gallagher, Ian
Jones, Andrew
Lawson, Daniel
Machine Learning
Methodology
Stability for dynamic network embeddings ensures that nodes behaving the same at different times receive the same embedding, allowing comparison of nodes in the network across time. We present attributed unfolded adjacency spectral embedding (AUASE), a stable unsupervised representation learning framework for dynamic networks in which nodes are attributed with time-varying covariate information. To establish stability, we prove uniform convergence to an associated latent position model. We quantify the benefits of our dynamic embedding by comparing with state-of-the-art network representation learning methods on four real attributed networks. To the best of our knowledge, AUASE is the only attributed dynamic embedding that satisfies stability guarantees without the need for ground truth labels, which we demonstrate provides significant improvements for link prediction and node classification.
title Unsupervised Attributed Dynamic Network Embedding with Stability Guarantees
topic Machine Learning
Methodology
url https://arxiv.org/abs/2503.02859