Abstract Deep neural networks (DNNs) provide superior performance on machine learning tasks such as image recognition, speech recognition, pattern analysis, and intrusion detection. However, an adversarial example, created by adding a little noise to an original sample, can cause misclassification by a DNN. This is a serious threat to the DNN because the added noise is not detected by the human eye. For example, if an attacker modifies a right-turn sign so that it misleads to the left, autonomous vehicles with the DNN will incorrectly classify the modified sign as pointing to the left, but a person will correctly classify the modified sign as pointing to the right. Studies are under way to defend against such adversarial examples. The existing method of defense against adversarial examples requires an additional process such as changing the classifier or modifying input data. In this paper, we propose a new method for detecting adversarial examples that does not invoke any additional process. The proposed scheme can detect adversarial examples by using a pattern feature of the classification scores of adversarial examples. We used MNIST and CIFAR10 as experimental datasets and Tensorflow as a machine learning library. The experimental results show that the proposed method can detect adversarial examples with success rates: 99.05% and 99.9% for the untargeted and targeted cases in MNIST, respectively, and 94.7% and 95.8% for the untargeted and targeted cases in CIFAR10, respectively.

Deep neural networks (DNNs) are widely used for image recognition, speech recognition, and other pattern analysis tasks. Despite the success of DNNs, these systems can be exploited by what is termed adversarial examples. An adversarial example, in which a small distortion is added to the input data, can be designed to be misclassified by the DNN while remaining undetected by humans or other systems. Such adversarial examples have been studied mainly in the image domain. Recently, however, studies on adversarial examples have been expanding into the voice domain. For example, when an adversarial example is applied to enemy wiretapping devices (victim classifiers) in a military environment, the enemy device will misinterpret the intended message. In such scenarios, it is necessary that friendly wiretapping devices (protected classifiers) should not be deceived. Therefore, the selective adversarial example concept can be useful in mixed situations, defined as situations in which there is both a classifier to be protected and a classifier to be attacked. In this paper, we propose a selective audio adversarial example with minimum distortion that will be misclassified as the target phrase by a victim classifier but correctly classified as the original phrase by a protected classifier. To generate such examples, a transformation is carried out to minimize the probability of incorrect classification by the protected classifier and that of correct classification by the victim classifier. We conducted experiments targeting the state-of-the-art DeepSpeech voice recognition model using Mozilla Common Voice datasets and the Tensorflow library. They showed that the proposed method can generate a selective audio adversarial example with a 91.67% attack success rate and 85.67% protected classifier accuracy.

Multihop relaying in WiMAX networks is considered an increasingly attractive technology for providing throughput enhancement, coverage extension, or both. In this paper, we consider deploying non-transparent relay stations (RSs) that transmit on the same carrier frequency as the base station (BS) for the purpose of coverage extension. Since it is possible that the RSs and BS are transmitting/receiving simultaneously to/from the subscriber stations (SSs) during the access zone period, it is challenging to schedule resources optimally for serving SSs in a fair manner. We propose an optimal scheduling scheme that preserves fairness among active SSs, and compare its performance with two different resource allocation schemes, namely, the overlapped and orthogonal in terms of cell throughput and outage performance. Our numerical results show that proposed scheme can achieve more throughput than the orthogonal scheme, while maintaining the fairness and lower outage performance of the orthogonal one.

Multihop relaying in WiMAX networks is considered a cost effective way to extend the coverage of the base stations, increase cell capacity, or both. In this paper, we consider deploying non-transparent relay stations for coverage extension and study the issue of fairness schemes in such networks. Since there are access and relay zones in both downlink and uplink frames and simultaneous transmissions are possible in the access zone, it is challenging to schedule resources optimally for serving subscriber stations in a fair manner. We evaluate the performance of well known fairness schemes such as max-min and proportional fairness in such networks. We also propose a new scheduling scheme, named subsection fairness that can achieve better throughput than traditional fairness schemes by maximizing bandwidth utilization. Our numerical results evaluate the performance of each scheduling scheme in terms of cell throughput and fairness.

WiMAX (IEEE 802.16) has emerged as a promising radio access technology for providing high speed broadband connectivity to subscribers over large geographic regions. New enhancements allow deployments of relay stations (RSs) that can extend the coverage of the base station (BS), increase cell capacity, or both. In this paper, we focus on deploying RSs for the purpose of coverage extension. We propose a simple scheduling scheme for serving the subscriber stations (SSs) in a fair manner and evaluate the performance of WiMAX networks with relays under this scheduling scheme. Through simulation and numerical analysis, we make several fundamental observations about cost-effective coverage in such networks.