Federated learning in omics studies hindered by batch effects, new paper finds

A new research paper published on arXiv explores the challenges of batch effects in federated learning for multi-center omics studies. The study demonstrates that uncorrected batch effects can significantly impair both unsupervised and supervised federated learning algorithms, including federated k-means clustering and federated random forest classification. To address this, the researchers introduced fedRBE, a novel federated implementation of limma's removeBatchEffect() method that utilizes secure multi-party computation for privacy-preserving batch-effect correction in distributed omics data. AI

IMPACT Highlights the need for robust privacy-preserving methods to ensure the reliability of federated learning in sensitive biomedical data analysis.

RANK_REASON Research paper published on arXiv detailing a new method for batch effect correction in federated learning for omics studies. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• federated k-means clustering
• federated learning
• federated random forest classification
• fedRBE
• limma
• Secure Multi-Party Computation
• Yuliya Burankova

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →