Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Implications of Decentralization in Blockchained Federated Learning: Evaluating the Impact of Model Staleness and Inconsistencies

Francesc Wilhelmi, Nima Afraz, Elia Guerra, Paolo Dini

TL;DR

This paper focuses on the effects that model staleness and inconsistencies, endorsed by blockchains' modus operandi, have on the training procedure held by FL devices asynchronously, and underscores the importance of properly designing blockchain systems based on the characteristics of the underlying FL application.

Abstract

Blockchain promises to enhance distributed machine learning (ML) approaches such as federated learning (FL) by providing further decentralization, security, immutability, and trust, which are key properties for enabling collaborative intelligence in next-generation applications. Nonetheless, the intrinsic decentralized operation of peer-to-peer (P2P) blockchain nodes leads to an uncharted setting for FL, whereby the concepts of FL round and global model become meaningless, as devices' synchronization is lost without the figure of a central orchestrating server. In this paper, we study the practical implications of outsourcing the orchestration of FL to a democratic setting such as in a blockchain. In particular, we focus on the effects that model staleness and inconsistencies, endorsed by blockchains' modus operandi, have on the training procedure held by FL devices asynchronously. Using simulation, we evaluate the blockchained FL operation by applying two different ML models (ranging from low to high complexity) on the well-known MNIST and CIFAR-10 datasets, respectively, and focus on the accuracy and timeliness of the solutions. Our results show the high impact of model inconsistencies on the accuracy of the models (up to a ~35% decrease in prediction accuracy), which underscores the importance of properly designing blockchain systems based on the characteristics of the underlying FL application.

The Implications of Decentralization in Blockchained Federated Learning: Evaluating the Impact of Model Staleness and Inconsistencies

TL;DR

This paper focuses on the effects that model staleness and inconsistencies, endorsed by blockchains' modus operandi, have on the training procedure held by FL devices asynchronously, and underscores the importance of properly designing blockchain systems based on the characteristics of the underlying FL application.

Abstract

Blockchain promises to enhance distributed machine learning (ML) approaches such as federated learning (FL) by providing further decentralization, security, immutability, and trust, which are key properties for enabling collaborative intelligence in next-generation applications. Nonetheless, the intrinsic decentralized operation of peer-to-peer (P2P) blockchain nodes leads to an uncharted setting for FL, whereby the concepts of FL round and global model become meaningless, as devices' synchronization is lost without the figure of a central orchestrating server. In this paper, we study the practical implications of outsourcing the orchestration of FL to a democratic setting such as in a blockchain. In particular, we focus on the effects that model staleness and inconsistencies, endorsed by blockchains' modus operandi, have on the training procedure held by FL devices asynchronously. Using simulation, we evaluate the blockchained FL operation by applying two different ML models (ranging from low to high complexity) on the well-known MNIST and CIFAR-10 datasets, respectively, and focus on the accuracy and timeliness of the solutions. Our results show the high impact of model inconsistencies on the accuracy of the models (up to a ~35% decrease in prediction accuracy), which underscores the importance of properly designing blockchain systems based on the characteristics of the underlying FL application.
Paper Structure (30 sections, 9 equations, 9 figures, 2 tables, 1 algorithm)

This paper contains 30 sections, 9 equations, 9 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. The decentralization of machine learning
  3. Challenges of blockchained federated learning
  4. Our contributions
  5. Structure of the document
  6. Related Work
  7. Decentralized federated learning
  8. Blockchained federated learning
  9. Blockchain and Federated Learning: Preliminaries
  10. Blockchain
  11. Federated learning
  12. Blockchained federated learning
  13. System Model
  14. Blockchained FL model
  15. Blockchain latency model
  16. ...and 15 more sections

Figures (9)

  • Figure 1: Overview of the procedures carried out by FL devices and blockchain nodes in blockchained FL.
  • Figure 2: Blockchain's protocol stack.
  • Figure 3: Blockchain structure and block information.
  • Figure 4: ML model training in Federated Averaging (FedAvg).
  • Figure 5: Blockchained FL logical components and interactions.
  • ...and 4 more figures