论文标题
DP $^2 $ -NILM：非侵入性负载监控的分布式和隐私保护框架
DP$^2$-NILM: A Distributed and Privacy-preserving Framework for Non-intrusive Load Monitoring
论文作者
论文摘要
Non-intrusive load monitoring (NILM), which usually utilizes machine learning methods and is effective in disaggregating smart meter readings from the household-level into appliance-level consumption, can help analyze electricity consumption behaviours of users and enable practical smart energy and smart grid applications.最近的研究提出了许多基于联邦深度学习（FL）的新型NILM框架。但是，在不同的基于FL的NILM应用程序方案中，缺乏探索实用性优化方案和隐私保护方案的全面研究。在本文中，我们首次尝试通过开发分布式和隐私的NILM（DP2-NILM）框架来进行基于FL的NILM，重点介绍公用事业优化和隐私保护，并在实用的NILM场景上进行基于现实世界智能仪表数据集的实际NILM场景进行比较实验。具体而言，在实用程序优化方案（即FedAvg和FedProx）中检查了两种替代联盟的学习策略。此外，DP2-NILM提供了不同级别的隐私保证，即联合学习的当地差异隐私学习和联盟的全球差异隐私学习。在三个现实世界数据集上进行了广泛的比较实验，以评估所提出的框架。
Non-intrusive load monitoring (NILM), which usually utilizes machine learning methods and is effective in disaggregating smart meter readings from the household-level into appliance-level consumption, can help analyze electricity consumption behaviours of users and enable practical smart energy and smart grid applications. Recent studies have proposed many novel NILM frameworks based on federated deep learning (FL). However, there lacks comprehensive research exploring the utility optimization schemes and the privacy-preserving schemes in different FL-based NILM application scenarios. In this paper, we make the first attempt to conduct FL-based NILM focusing on both the utility optimization and the privacy-preserving by developing a distributed and privacy-preserving NILM (DP2-NILM) framework and carrying out comparative experiments on practical NILM scenarios based on real-world smart meter datasets. Specifically, two alternative federated learning strategies are examined in the utility optimization schemes, i.e., the FedAvg and the FedProx. Moreover, different levels of privacy guarantees, i.e., the local differential privacy federated learning and the global differential privacy federated learning are provided in the DP2-NILM. Extensive comparison experiments are conducted on three real-world datasets to evaluate the proposed framework.
