Abstract
Understanding residential household characteristics is crucial for retailers to provide customers personalized services. Current methods infer household characteristics from smart meter data in a centralized manner that requires the data of all retailers to be gathered together for model training. This may raise privacy concerns since the privacy-sensitive data are owned by different retailers, and they may be unwilling to share the raw data. This paper proposes a federated learning (FL) based deep learning model to identify household characteristics. A hybrid model combining the convolutional neural network and long short-term neural network is designed to learn spatial-temporal features from load profiles. It is implemented in a decentralized manner based on the FL framework. To improve the training speed and accuracy, an asynchronous stochastic gradient descent with delay compensation method is proposed to update the global model parameters. Comprehensive experiments are conducted to verify the effectiveness of the proposed method.
Funding Information
  • Australia–China Science and Research Fund Joint Research Center for Energy Informatics and Demand Response Technologies
  • Australian Research Council (ARC) through A Unified Framework for Resource Management in Edge-Cloud Data Centers (DP200103494)

This publication has 32 references indexed in Scilit: