Federated Causal Discovery Across Heterogeneous Datasets under Latent Confounding
#federated learning #causal discovery #latent confounding #heterogeneous data #privacy preservation #decentralized datasets #causal inference
📌 Key Takeaways
- Federated causal discovery enables causal inference across decentralized datasets without sharing raw data.
- The method addresses challenges of heterogeneous data distributions and latent confounding variables.
- It preserves privacy by keeping data local while learning global causal structures.
- The approach improves accuracy in identifying causal relationships compared to non-federated methods.
📖 Full Retelling
arXiv:2603.05149v1 Announce Type: cross
Abstract: Causal discovery across multiple datasets is often constrained by data privacy regulations and cross-site heterogeneity, limiting the use of conventional methods that require a single, centralized dataset. To address these challenges, we introduce fedCI, a federated conditional independence test that rigorously handles heterogeneous datasets with non-identical sets of variables, site-specific effects, and mixed variable types, including continuo
🏷️ Themes
Causal Discovery, Federated Learning, Data Privacy
Entity Intersection Graph
No entity connections available yet for this article.
Original Source
--> Computer Science > Machine Learning arXiv:2603.05149 [Submitted on 5 Mar 2026] Title: Federated Causal Discovery Across Heterogeneous Datasets under Latent Confounding Authors: Maximilian Hahn , Alina Zajak , Dominik Heider , Adèle Helena Ribeiro View a PDF of the paper titled Federated Causal Discovery Across Heterogeneous Datasets under Latent Confounding, by Maximilian Hahn and 3 other authors View PDF Abstract: Causal discovery across multiple datasets is often constrained by data privacy regulations and cross-site heterogeneity, limiting the use of conventional methods that require a single, centralized dataset. To address these challenges, we introduce fedCI, a federated conditional independence test that rigorously handles heterogeneous datasets with non-identical sets of variables, site-specific effects, and mixed variable types, including continuous, ordinal, binary, and categorical variables. At its core, fedCI uses a federated Iteratively Reweighted Least Squares procedure to estimate the parameters of generalized linear models underlying likelihood-ratio tests for conditional independence. Building on this, we develop fedCI-IOD, a federated extension of the Integration of Overlapping Datasets algorithm, that replaces its meta-analysis strategy and enables, for the fist time, federated causal discovery under latent confounding across distributed and heterogeneous datasets. By aggregating evidence federatively, fedCI-IOD not only preserves privacy but also substantially enhances statistical power, achieving performance comparable to fully pooled analyses and mitigating artifacts from low local sample sizes. Our tools are publicly available as the fedCI Python package, a privacy-preserving R implementation of IOD, and a web application for the fedCI-IOD pipeline, providing versatile, user-friendly solutions for federated conditional independence testing and causal discovery. Subjects: Machine Learning (cs.LG) ; Artificial Intelligence (cs.AI); Machine L...
Read full article at source