Biblio

Deep learning is becoming a basis of decision making systems in many application domains, such as autonomous vehicles, health systems, etc., where the risk of misclassification can lead to serious consequences. It is necessary to know to which extent are Deep Neural Networks (DNNs) robust against various types of adversarial conditions. In this paper, we experimentally evaluate DNNs implemented in embedded device by using laser fault injection, a physical attack technique that is mostly used in security and reliability communities to test robustness of various systems. We show practical results on four activation functions, ReLu, softmax, sigmoid, and tanh. Our results point out the misclassification possibilities for DNNs achieved by injecting faults into the hidden layers of the network. We evaluate DNNs by using several different attack strategies to show which are the most efficient in terms of misclassification success rates. Outcomes of this work should be taken into account when deploying devices running DNNs in environments where malicious attacker could tamper with the environmental parameters that would bring the device into unstable conditions. resulting into faults.

In multicarrier direct modulation direct detection systems, interaction between laser chirp and fiber group velocity dispersion induces subcarrier-to-subcarrier intermixing interferences (SSII) after detection. Such SSII become a major impairment in orthogonal frequency division multiplexing-based access systems, where a high modulation index, leading to large chirp, is required to maximize the system power budget. In this letter, we present and experimentally verify an analytical formulation to predict the level of signal and SSII and estimate the signal to noise ratio of each subcarrier, enabling improved bit-and-power loading and subcarrier attribution. The reported model is compact, and only requires the knowledge of basic link characteristics and laser parameters that can easily be measured.