Causality-Driven Disentangled Representation Learning in Multiplex Graphs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nasiri, Saba, Aviyente, Selin, Thanou, Dorina
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917361130405888
author Nasiri, Saba
Aviyente, Selin
Thanou, Dorina
author_facet Nasiri, Saba
Aviyente, Selin
Thanou, Dorina
contents Learning representations from multiplex graphs, i.e., multi-layer networks where nodes interact through multiple relation types, is challenging due to the entanglement of shared (common) and layer-specific (private) information, which limits generalization and interpretability. In this work, we introduce a causal inference-based framework that disentangles common and private components in a self-supervised manner. CaDeM jointly (i) aligns shared embeddings across layers, (ii) enforces private embeddings to capture layer-specific signals, and (iii) applies backdoor adjustment to ensure that the common embeddings capture only global information while being separated from the private representations. Experiments on synthetic and real-world datasets demonstrate consistent improvements over existing baselines, highlighting the effectiveness of our approach for robust and interpretable multiplex graph representation learning.
format Preprint
id arxiv_https___arxiv_org_abs_2603_24105
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Causality-Driven Disentangled Representation Learning in Multiplex Graphs
Nasiri, Saba
Aviyente, Selin
Thanou, Dorina
Machine Learning
Social and Information Networks
Learning representations from multiplex graphs, i.e., multi-layer networks where nodes interact through multiple relation types, is challenging due to the entanglement of shared (common) and layer-specific (private) information, which limits generalization and interpretability. In this work, we introduce a causal inference-based framework that disentangles common and private components in a self-supervised manner. CaDeM jointly (i) aligns shared embeddings across layers, (ii) enforces private embeddings to capture layer-specific signals, and (iii) applies backdoor adjustment to ensure that the common embeddings capture only global information while being separated from the private representations. Experiments on synthetic and real-world datasets demonstrate consistent improvements over existing baselines, highlighting the effectiveness of our approach for robust and interpretable multiplex graph representation learning.
title Causality-Driven Disentangled Representation Learning in Multiplex Graphs
topic Machine Learning
Social and Information Networks
url https://arxiv.org/abs/2603.24105