An Ontological Knowledge Base of Poisoning Attacks on Deep Neural Networks

Abstract

Deep neural networks (DNNs) have successfully delivered cutting-edge performance in several fields. With the broader deployment of DNN models on critical applications, the security of DNNs has become an active and yet nascent area. Attacks against DNNs can have catastrophic results, according to recent studies. Poisoning attacks, including backdoor attacks and Trojan attacks, are one of the growing threats against DNNs. Having a wide-angle view of these evolving threats is essential to better understand the security issues. In this regard, creating a semantic model and a knowledge graph for poisoning attacks can reveal the relationships between attacks across intricate data to enhance the security knowledge landscape. In this paper, we propose a DNN poisoning attack ontology (DNNPAO) that would enhance knowledge sharing and enable further advancements in the field. To do so, we have performed a systematic review of the relevant literature to identify the current state. We collected 28,469 papers from the IEEE, ScienceDirect, Web of Science, and Scopus databases, and from these papers, 712 research papers were screened in a rigorous process, and 55 poisoning attacks in DNNs were identified and classified. We extracted a taxonomy of the poisoning attacks as a scheme to develop DNNPAO. Subsequently, we used DNNPAO as a framework by which to create a knowledge base. Our findings open new lines of research within the field of AI security.

AI explainability and governance in smart energy systems: A review

“…Specific fault detection-related challenges include the management and storage issues arising due to the use of multiple data sources (solar or wind power forecasting and related faults using numerical and image data), as opposed to a single data source, for fault detection (Landwehr et al, 2022), the influence of measurement noise on fault prediction performance (Sun et al, 2021), privacy issues in faultdiagnosis and examination of security and stability-sensitive scenarios (Ardito et al, 2022), and low accuracies of AI algorithms in fault detection, diagnosis, and prediction (Wu et al, 2022). Another Frontiers in Energy Research frontiersin.org increasingly important area is the security of ML and DL software (Altoub et al, 2022). The challenges in this area include, among others, data poisoning attacks and the performance of related solutions (Bhattacharjee et al, 2022), and anomaly detection methods for smart grid meter security against poisoning attacks (Bhattacharjee et al, 2021).…”

Understanding Local Government Cybersecurity Policy: A Concept Map and Framework

“…The smart city movement, which significantly enhances urban digital capabilities, also increases our cities' vulnerability to cybersecurity threats [1][2][3][4][5]. In the age of smart cities and digital transformation, local governments (LGs) face increasing cybersecurity threats due to storing and managing a vast amount of sensitive information, including residents' data and critical infrastructure details [6][7][8].…”