P4L: Privacy Preserving Peer-to-Peer Learning for Infrastructureless Setups

TL;DR

P4L introduces a peer-to-peer, privacy-preserving collaborative learning system that eliminates the need for centralized infrastructure, using cryptography and fault-tolerant mechanisms, with minimal performance overhead.

Contribution

The paper presents P4L, a novel infrastructureless peer-to-peer learning system that ensures privacy and robustness without relying on differential privacy or centralized servers.

Findings

P4L achieves competitive accuracy compared to baseline models.

The system is resilient to poisoning attacks.

Performance overhead and power consumption are minimal.

Abstract

Distributed (or Federated) learning enables users to train machine learning models on their very own devices, while they share only the gradients of their models usually in a differentially private way (utility loss). Although such a strategy provides better privacy guarantees than the traditional centralized approach, it requires users to blindly trust a centralized infrastructure that may also become a bottleneck with the increasing number of users. In this paper, we design and implement P4L: a privacy preserving peer-to-peer learning system for users to participate in an asynchronous, collaborative learning scheme without requiring any sort of infrastructure or relying on differential privacy. Our design uses strong cryptographic primitives to preserve both the confidentiality and utility of the shared gradients, a set of peer-to-peer mechanisms for fault tolerance and user churn, proximity and cross device communications. Extensive simulations under different network settings and ML scenarios for three real-life datasets show that P4L provides competitive performance to baselines, while it is resilient to different poisoning attacks. We implement P4L and experimental results show that the performance overhead and power consumption is minimal (less than 3mAh of discharge).