This is the authors' abstract. We don't add key points for this paper.
Non-Intrusive Load Monitoring (NILM) refers to as the technology of identifying the operation status and power consumption of individual electrical devices (typically household appliances) from an aggregated smart meter reading profile. This paper proposes a generative, transferrable NILM system called “GT-NILM”, which can provide accurate NILM services to stakeholders (e.g., grid operators and building managers) for understanding the operational information of individual appliances of residential households while sufficiently preserve the households energy data privacy. The proposed system is backboned a conditional diffusion model and a convolutional neural network, which are designed for identifying the appliances operating power waveform characteristics and ON/OFF status from smart meter readings, respectively. This dual-model design enables fast model training while maintaining high NILM accuracy, with performance validated on REDD and UK-DALE datasets. In standard within-dataset evaluations, GT-NILM achieves comparable performance to state-of-the-art NILM models, while in a cross-dataset transfer learning experiment, it outperforms other state-of-the-art methods up to 10.95% on the F1-score metric and 36.31% on the mean absolute error metric.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Luo et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: