New federated clustering method enhances privacy and efficiency

Researchers have developed a novel one-shot federated clustering approach called Evolving Federated Density Peak Clustering (eFDPC). This method combines evolving fuzzy data-stream clustering with density peak clustering in a federated setting, allowing clients to independently evolve micro-clusters from private data. To minimize communication, only selected peaks are transmitted to the server for global selection. A shuffler-based layer is integrated to anonymize client contributions before reconstruction, enhancing privacy. Experiments indicate that eFDPC surpasses existing federated clustering baselines in one-shot scenarios and effectively handles high-dimensional, non-independent and identically distributed data. AI

IMPACT This new federated clustering method could improve privacy and efficiency in distributed machine learning tasks.

RANK_REASON The item is an academic paper detailing a new method for federated clustering. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• eFDPC
• eFDPC-G
• eFDPC-P
• Evolving Federated Density Peak Clustering
• Federated Density Peak Clustering
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Litmaps
• ScienceCast
• scite Smart Citations
• Shuffler-Based Anonymization

• paper
• other

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