Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Spiking Neural Networks in Vertical Federated Learning: Performance Trade-offs

Maryam Abbasihafshejani, Anindya Maiti, Murtuza Jadliwala

TL;DR

The paper addresses energy-efficient privacy-preserving learning in Vertical Federated Learning by integrating Spiking Neural Networks (SNNs) and evaluating two architectures: with and without model splitting. It adapts SNN models (VGG9 and ResNet18) to CIFAR-10/100 within a VFL framework, using rate Poisson encoding and a 32-time-step regime, and analyzes accuracy and energy against ANN baselines. Results show SNN-based VFL achieves comparable accuracy to ANN-based VFL while dramatically reducing energy consumption (for example, around a 34.7x energy improvement on CIFAR-10 with VGG9), albeit with longer training times. This work demonstrates the practical viability of neuromorphic models for privacy-preserving edge learning and informs trade-offs between energy efficiency and training speed for distributed, feature-partitioned data settings, guiding future hardware-aware optimizations and model variants.

Abstract

Federated machine learning enables model training across multiple clients while maintaining data privacy. Vertical Federated Learning (VFL) specifically deals with instances where the clients have different feature sets of the same samples. As federated learning models aim to improve efficiency and adaptability, innovative neural network architectures like Spiking Neural Networks (SNNs) are being leveraged to enable fast and accurate processing at the edge. SNNs, known for their efficiency over Artificial Neural Networks (ANNs), have not been analyzed for their applicability in VFL, thus far. In this paper, we investigate the benefits and trade-offs of using SNN models in a vertical federated learning setting. We implement two different federated learning architectures -- with model splitting and without model splitting -- that have different privacy and performance implications. We evaluate the setup using CIFAR-10 and CIFAR-100 benchmark datasets along with SNN implementations of VGG9 and ResNET classification models. Comparative evaluations demonstrate that the accuracy of SNN models is comparable to that of traditional ANNs for VFL applications, albeit significantly more energy efficient.

Spiking Neural Networks in Vertical Federated Learning: Performance Trade-offs

TL;DR

The paper addresses energy-efficient privacy-preserving learning in Vertical Federated Learning by integrating Spiking Neural Networks (SNNs) and evaluating two architectures: with and without model splitting. It adapts SNN models (VGG9 and ResNet18) to CIFAR-10/100 within a VFL framework, using rate Poisson encoding and a 32-time-step regime, and analyzes accuracy and energy against ANN baselines. Results show SNN-based VFL achieves comparable accuracy to ANN-based VFL while dramatically reducing energy consumption (for example, around a 34.7x energy improvement on CIFAR-10 with VGG9), albeit with longer training times. This work demonstrates the practical viability of neuromorphic models for privacy-preserving edge learning and informs trade-offs between energy efficiency and training speed for distributed, feature-partitioned data settings, guiding future hardware-aware optimizations and model variants.

Abstract

Federated machine learning enables model training across multiple clients while maintaining data privacy. Vertical Federated Learning (VFL) specifically deals with instances where the clients have different feature sets of the same samples. As federated learning models aim to improve efficiency and adaptability, innovative neural network architectures like Spiking Neural Networks (SNNs) are being leveraged to enable fast and accurate processing at the edge. SNNs, known for their efficiency over Artificial Neural Networks (ANNs), have not been analyzed for their applicability in VFL, thus far. In this paper, we investigate the benefits and trade-offs of using SNN models in a vertical federated learning setting. We implement two different federated learning architectures -- with model splitting and without model splitting -- that have different privacy and performance implications. We evaluate the setup using CIFAR-10 and CIFAR-100 benchmark datasets along with SNN implementations of VGG9 and ResNET classification models. Comparative evaluations demonstrate that the accuracy of SNN models is comparable to that of traditional ANNs for VFL applications, albeit significantly more energy efficient.
Paper Structure (27 sections, 8 equations, 5 figures, 3 tables, 2 algorithms)

This paper contains 27 sections, 8 equations, 5 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Spiking Neural Networks (SNNs)
  4. Input Encoding
  5. SNN Models
  6. Output Decoding
  7. Batch Normalization Through Time (BNTT)
  8. Federated Learning
  9. Horizontal Federated Learning
  10. Vertical Federated Learning
  11. SNN and Federated Learning
  12. Vertical Federated Training of Spiking Neural Networks
  13. VFL-SNN with Model Splitting
  14. VFL-SNN without Model Splitting
  15. Evaluation
  16. ...and 12 more sections

Figures (5)

  • Figure 1: VFL (A) without model splitting, and (B) with model splitting.
  • Figure 2: Comparison of CIFAR-10 and CIFAR-100 classification accuracy across different numbers of clients.
  • Figure 3: Variations in accuracy with increasing time steps.
  • Figure 4: Estimated training energy across layers for SNN (VFL-SNN ) and ANN VGG9 model on CIFAR-10 dataset.
  • Figure 5: Training time comparison across models.