A NILM method using multimodal image load signatures with self-supervised pretraining and EWC continual learning reports 92.5% accuracy, but the continual learning component is not experimentally evaluated.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Non-Intrusive Load Monitoring Based on Image Load Signatures and Continual Learning
A NILM method using multimodal image load signatures with self-supervised pretraining and EWC continual learning reports 92.5% accuracy, but the continual learning component is not experimentally evaluated.