Detecting One-Pixel Attacks Using Variational Autoencoders

Alatalo, Janne; Sipola, Tuomo; Kokkonen, Tero

doi:10.1007/978-3-031-04826-5_60

Cited by 5 publications

(1 citation statement)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Like all IT systems, the food supply chain is vulnerable to cyber security threats, e.g., in the areas of smart farming and cyber-physical systems [9]. Cyber attacks against food supply chains can be devastating.…”

Section: Cyber Exercisesmentioning

confidence: 99%

Digital Twin of Food Supply Chain for Cyber Exercises

Sipola,

Kokkonen,

Puura

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

The food supply chain is a critical part of modern societies. As with other facets of life, it is thoroughly digitalized, and uses network connections. Consequently, the cyber security of the supply chain becomes a major concern as new threats emerge. Cyber ranges can be used to prepare for such cyber security threats by creating realistic scenarios mimicking real-world systems and setups. Organizations can participate in cyber security training and exercises that present them with these scenarios. Cyber ranges can also be used efficiently for research and development activities, because cyber ranges are realistic environments and can be used for the generation of realistic data. The aim of this study is to describe a digital twin of the food supply chain built for cyber range-based cyber security exercises. The digital twin mirrors the real-world situation with sufficient detail, as required by the cyber exercise. This research uses the design science methodology, which describes the construction and evaluation of the proposed system. The study explains the general capabilities of the food supply chain digital twin and its use in the cyber range environment. Different parts of the supply chain are implemented as Node.js services that run on the Realistic Global Cyber Environment (RGCE) platform. The flow of ingredients and products is simulated using an apparatus model and message queues. The digital twin was demonstrated in a real live cyber exercise. The results indicate that the apparatus approach was a scalable and realistic enough way to implement the digital twin. The main limitations of the implemented system are the implementation on one specific platform, and the need for more feedback from multiple exercises. Creation of a digital twin enables the use of cyber ranges to train organizations related to the food supply chain.

show abstract

Section: Cyber Exercisesmentioning

confidence: 99%

Digital Twin of Food Supply Chain for Cyber Exercises

Sipola,

Kokkonen,

Puura

et al. 2023

Applied Sciences

View full text Add to dashboard Cite

show abstract

Medical images under tampering

Tsai,

Lin

2024

Multimed Tools Appl

View full text Add to dashboard Cite

Attacks on deep learning models are a constant threat in the world today. As more deep learning models and artificial intelligence (AI) are being implemented across different industries, the likelihood of them being attacked increases dramatically. In this context, the medical domain is of the greatest concern because an erroneous decision made by AI could have a catastrophic outcome and even lead to death. Therefore, a systematic procedure is built in this study to determine how well these medical images can resist a specific adversarial attack, i.e. a one-pixel attack. This may not be the strongest attack, but it is simple and effective, and it could occur by accident or an equipment malfunction. The results of the experiment show that it is difficult for medical images to survive a one-pixel attack.

show abstract