FedPS framework enables privacy-preserving data preprocessing for federated learning

Researchers have introduced FedPS, a novel framework designed for federated data preprocessing. This system enables collaborative machine learning model training across multiple parties without the need to share raw data. FedPS utilizes data-sketching techniques to efficiently summarize local datasets, preserving key statistical information. The framework supports federated algorithms for feature scaling, encoding, discretization, and missing-value imputation, extending preprocessing models like Bayesian Linear Regression to both horizontal and vertical federated learning settings. AI

IMPACT Enhances privacy and efficiency in collaborative machine learning model training by addressing critical data preprocessing challenges.

RANK_REASON This is a research paper detailing a new framework for federated learning preprocessing. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• arXiv
• Bayesian linear regression
• federated learning
• Xuefeng Xu

• paper
• infra

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