A novel data enhancement approach to DAG learning with small data samples

Huang, Xiaoling; Guo, Xianjie; Li, Yuling; Yu, Kui

doi:10.1007/s10489-023-04999-2

Cited by 4 publications

(1 citation statement)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we propose a novel federated CSL method, called FedCSL, to tackle the two challenges above and make the following main contributions. A causal structure over X is often represented using a causal directed acyclic graph (DAG) (Huang et al 2023b). In a causal DAG, if there is a direct edge X i → X j (i, j ∈ {1, 2, ..., d}), X i is a direct cause (parent) of X j , and X j is a direct effect (child) of X i (Wu et al 2022).…”

Section: Contributionsmentioning

confidence: 99%

FedCSL: A Scalable and Accurate Approach to Federated Causal Structure Learning

Guo,

Yu,

Liu

et al. 2024

AAAI

View full text Add to dashboard Cite

As an emerging research direction, federated causal structure learning (CSL) aims at learning causal relationships from decentralized data across multiple clients while preserving data privacy. Existing federated CSL algorithms suffer from scalability and accuracy issues, since they require computationally expensive CSL algorithms to be executed at each client. Furthermore, in real-world scenarios, the number of samples held by each client varies significantly, and existing methods still assign equal weights to the learned structural information from each client, which severely harms the learning accuracy of those methods. To address these two limitations, we propose FedCSL, a scalable and accurate method for federated CSL. Specifically, FedCSL consists of two novel strategies: (1) a federated local-to-global learning strategy that enables FedCSL to scale to high-dimensional data for tackling the scalability issue, and (2) a novel weighted aggregation strategy that does not rely on any complex encryption techniques while preserving data privacy for tackling the accuracy issue. Extensive experiments on benchmark datasets, high-dimensional synthetic datasets and a real-world dataset verify the efficacy of the proposed FedCSL method. The source code is available at https://github.com/Xianjie-Guo/FedCSL.

show abstract

Section: Contributionsmentioning

confidence: 99%