Causal representation learning from network data

2025-09-01

AI Summary

Addressing causal representation learning under non-i.i.d. conditions in network-structured data, this paper proposes GraCE-VAE—the first framework to extend causal identifiability theory from the i.i.d. setting to structured graph data. GraCE-VAE integrates graph neural networks with a difference-driven variational autoencoder to jointly model observational and interventional data, enabling simultaneous recovery of the latent causal graph structure and intervention effects. Its core innovation lies in explicitly leveraging network topology as a structural prior to enhance causal disentanglement. Extensive experiments on three gene perturbation datasets demonstrate that GraCE-VAE significantly outperforms existing baselines, achieving state-of-the-art performance in both causal graph reconstruction and intervention effect estimation. These results empirically validate the critical role of network structure in improving causal representation learning.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies

Abstract

Causal disentanglement from soft interventions is identifiable under the assumptions of linear interventional faithfulness and availability of both observational and interventional data. Previous research has looked into this problem from the perspective of i.i.d. data. Here, we develop a framework, GraCE-VAE, for non-i.i.d. settings, in which structured context in the form of network data is available. GraCE-VAE integrates discrepancy-based variational autoencoders with graph neural networks to jointly recover the true latent causal graph and intervention effects. We show that the theoretical results of identifiability from i.i.d. data hold in our setup. We also empirically evaluate GraCE-VAE against state-of-the-art baselines on three genetic perturbation datasets to demonstrate the impact of leveraging structured context for causal disentanglement.

Problem

• Causal disentanglement from soft interventions under network data
• Recovering latent causal graph and intervention effects jointly
• Leveraging structured context for improved causal representation learning

Innovation

• Graph neural networks integrate structured context data
• Variational autoencoders recover latent causal graphs
• Framework handles non-i.i.d. network data settings