A4 Article in conference proceedings
Influence of attack vectors on generic artificial intelligence : Assisted smart building feedback loop system (2020)
Vähäkainu, P., Lehto, M., & Kariluoto, A. (2020). Influence of attack vectors on generic artificial intelligence : Assisted smart building feedback loop system. In T. Eze, L. Speakman, & C. Onwubiko (Eds.), ECCWS 2020 : Proceedings of the 19th European Conference on Cyber Warfare and Security (pp. 371-379). Academic Conferences International. Proceedings of the European conference on information warfare and security. https://doi.org/10.34190/EWS.20.046
JYU authors or editors
Publication details
All authors or editors: Vähäkainu, Petri; Lehto, Martti; Kariluoto, Antti
Parent publication: ECCWS 2020 : Proceedings of the 19th European Conference on Cyber Warfare and Security
Parent publication editors: Eze, Thaddeus; Speakman, Lee; Onwubiko, Cyril
Conference:
- European Conference on Cyber Warfare and Security
Place and date of conference: University of Chester, UK (virtual conference), 25-26.6.2020
ISBN: 978-1-912764-61-7
eISBN: 978-1-912764-62-4
Journal or series: Proceedings of the European conference on information warfare and security
ISSN: 2048-8602
eISSN: 2048-8610
Publication year: 2020
Pages range: 371-379
Publisher: Academic Conferences International
Publication country: United Kingdom
Publication language: English
DOI: https://doi.org/10.34190/EWS.20.046
Publication open access: Not open
Publication channel open access:
Abstract
A Cyber-Physical system utilizes real-world measurements collected via IoT sensors. Using data on measurements and outside weather, it is possible to train a machine-learning model that can automatically adjust HVAC and other heating devices of smart buildings. With traditional systems, it is not possible to attain the required energy efficiency values. Instead, artificial intelligence could minimize these values. However, in case the perpetrator can influence the AI governing the feedback system, the cost may increase significantly. For example, the cost of power spikes has a long payback, which can be multiple years. In this article, we define the concepts of cybersecurity and machine learning, do a literary review on machine learning models, recognition of the relevant attack vectors to a feedback system of a smart building and countering of those attacks, and discuss the generic feedback system implementation.
Keywords: intelligent systems; smart houses; cyber security; cyber attacks; artificial intelligence; machine learning
Free keywords: adversarial attacks; artificial intelligence; artificial intelligence-based applications; attack vectors; cloud service; data platform
Contributing organizations
Ministry reporting: Yes
Reporting Year: 2020
JUFO rating: 1