KIT | KIT-Bibliothek | Impressum | Datenschutz
Open Access Logo
§
Verlagsausgabe
DOI: 10.5445/IR/1000086898
Veröffentlicht am 24.10.2018
Originalveröffentlichung
DOI: 10.1186/s42162-018-0038-y

Enhancing neural non-intrusive load monitoring with generative adversarial networks

Bao, Kaibin; Ibrahimov, Kanan; Wagner, Martin; Schmeck, Hartmut

Abstract (englisch):
The application of Deep Learning methodologies to Non-Intrusive Load Monitoring (NILM) gave rise to a new family of Neural NILM approaches which increasingly outperform traditional NILM approaches. In this extended abstract describing our ongoing research, we analyze recent Neural NILM approaches and our findings imply that these approaches have difficulties in generating valid, reasonably-shaped appliance load profiles. We propose to enhance Neural NILM approaches with appliance load sequence generators trained with a Generative Adversarial Network to mitigate the described problem. The preliminary results of our experiments with Generative Adversarial Networks show the potential of the approach, albeit there is no strong evidence yet that this approach outperforms the examined end-to-end-trained Neural NILM approaches. In the progress of our investigations, we generalize energy-based NILM performance metrics and establish the complete classification confusion matrix based on the estimated energy in appliance load profiles. This enables the adaption of all known classification scores to their energy-based counterparts.


Zugehörige Institution(en) am KIT Institut für Angewandte Informatik und Formale Beschreibungsverfahren (AIFB)
Kompetenzzentrum für angewandte Sicherheitstechnologie (KASTEL)
Publikationstyp Zeitschriftenaufsatz
Jahr 2018
Sprache Englisch
Identifikator ISSN: 2520-8942
URN: urn:nbn:de:swb:90-868989
KITopen ID: 1000086898
Erschienen in Energy Informatics
Band S1
Heft 18
Seiten 295-302
Projektinformation KASTEL_SVI (BMBF, 16KIS0521)
Bemerkung zur Veröffentlichung Proceedings of the 7th DACH+ Conference on Energy Informatics, Oldenburg, Germany. 11-12 October 2018
Vorab online veröffentlicht am 10.10.2018
Schlagworte Non-intrusive load monitoring; Generative adversarial networks; Neural NILM; Generative modeling; Deep Learning
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft KITopen Landing Page