Privacy-Preserving Load Forecasting via Personalized Model Obfuscation

TL;DR

This paper introduces PPFL, a federated learning approach with personalized layers and differential privacy to enhance load forecasting accuracy while protecting user privacy in smart meter data.

Contribution

It proposes a novel federated learning framework with personalization and privacy mechanisms specifically designed for load forecasting on heterogeneous smart meter data.

Findings

PPFL improves load forecasting accuracy under privacy constraints.

Differential privacy effectively prevents data leakage from shared models.

Personalization layers enhance model performance on local data.

Abstract

The widespread adoption of smart meters provides access to detailed and localized load consumption data, suitable for training building-level load forecasting models. To mitigate privacy concerns stemming from model-induced data leakage, federated learning (FL) has been proposed. This paper addresses the performance challenges of short-term load forecasting models trained with FL on heterogeneous data, emphasizing privacy preservation through model obfuscation. Our proposed algorithm, Privacy Preserving Federated Learning (PPFL), incorporates personalization layers for localized training at each smart meter. Additionally, we employ a differentially private mechanism to safeguard against data leakage from shared layers. Simulations on the NREL ComStock dataset corroborate the effectiveness of our approach.