WPFed: Web-based Personalized Federation for Decentralized Systems
This addresses the problem of effective and secure decentralized collaborative learning for web environments with data privacy concerns, though it appears incremental as it builds on existing federated learning approaches.
The paper tackles the challenge of optimal neighbor selection in decentralized learning for web-based applications, introducing WPFed, a framework that uses dynamic communication graphs and weighted neighbor selection to improve learning outcomes and system robustness, with experiments on real-world datasets showing significant improvements over traditional federated learning methods.
Decentralized learning has become crucial for collaborative model training in environments where data privacy and trust are paramount. In web-based applications, clients are liberated from traditional fixed network topologies, enabling the establishment of arbitrary peer-to-peer (P2P) connections. While this flexibility is highly promising, it introduces a fundamental challenge: the optimal selection of neighbors to ensure effective collaboration. To address this, we introduce WPFed, a fully decentralized, web-based learning framework designed to enable globally optimal neighbor selection. WPFed employs a dynamic communication graph and a weighted neighbor selection mechanism. By assessing inter-client similarity through Locality-Sensitive Hashing (LSH) and evaluating model quality based on peer rankings, WPFed enables clients to identify personalized optimal neighbors on a global scale while preserving data privacy. To enhance security and deter malicious behavior, WPFed integrates verification mechanisms for both LSH codes and performance rankings, leveraging blockchain-driven announcements to ensure transparency and verifiability. Through extensive experiments on multiple real-world datasets, we demonstrate that WPFed significantly improves learning outcomes and system robustness compared to traditional federated learning methods. Our findings highlight WPFed's potential to facilitate effective and secure decentralized collaborative learning across diverse and interconnected web environments.