No abstract

No abstract

Cardiac implantable electronic devices (CIED) are vulnerable to radio frequency (RF) cyber-attacks. Besides, CIED communicate with medical equipment whose telemetry capabilities and IP connectivity are creating new entry points that may be used by attackers. Therefore, it remains crucial to perform a cybersecurity risk assessment of CIED and the systems they rely on to determine the gravity of threats, address the riskiest ones on a priority basis, and develop effective risk management plans. In this study, we carry out such risk assessment according to the ISO/IEC 27005 standard and the NIST SP 800-30 guide. We employed a threat-oriented analytical approach and divided the analysis into three parts, an actor-based analysis to determine the impact of the attacks, a scenario-based analysis to measure the probability of occurrence of threats, and a combined analysis to identify the riskiest attack outcomes. The results show that vulnerabilities on the RF interface of CIED represent an acceptable risk, whereas the network and Internet connectivity of the systems they rely on represent an important potential risk. Further analysis reveals that the damages of these cyber-attacks could spread further to affect manufacturers through intellectual property theft or physicians by affecting their reputation.

No abstract

Deep learning models are vulnerable to adversarial examples which are input samples modified in order to maximize the error on the system. We introduce Spartan Networks, resistant deep neural networks that do not require input preprocessing nor adversarial training. These networks have an adversarial layer designed to discard some information of the network, thus forcing the system to focus on relevant input. This is done using a new activation function to discard data. The added layer trains the neural network to filter-out usually-irrelevant parts of its input. Our performance evaluation shows that Spartan Networks have a slightly lower precision but report a higher robustness under attack when compared to unprotected models. Results of this study of Adversarial AI as a new attack vector are based on tests conducted on the MNIST dataset. KEYWORDSArtificial Intelligence, Cybersecurity, Adversarial AI Potentially any deep learning model can be vulnerable to this kind of attack, which is hard to detect, prevent, and whose impact will only grow in the upcoming years.