Search

Mobile devices such as sensors are used to connect to the Internet and provide services to users. Web services are vulnerable to automated attacks, which can restrict mobile devices from accessing websites. To prevent such automated attacks, CAPTCHAs are widely used as a security solution. However, when a high level of distortion has been applied to a CAPTCHA to make it resistant to automated attacks, the CAPTCHA becomes difficult for a human to recognize. In this work, we propose a method for generating a CAPTCHA image that will resist recognition by machines while maintaining its recognizability to humans. The method utilizes the style transfer method, and creates a new image, called a style-plugged-CAPTCHA image, by incorporating the styles of other images while keeping the content of the original CAPTCHA. In our experiment, we used the TensorFlow machine learning library and six CAPTCHA datasets in use on actual websites. The experimental results show that the proposed scheme reduces the rate of recognition by the DeCAPTCHA system to 3.5% and 3.2% using one style image and two style images, respectively, while maintaining recognizability by humans.

Studies related to pattern recognition and visualization using computer technology have been introduced. In particular, deep neural networks (DNNs) provide good performance for image, speech, and pattern recognition. However, a poisoning attack is a serious threat to a DNN's security. A poisoning attack reduces the accuracy of a DNN by adding malicious training data during the training process. In some situations, it may be necessary to drop a specifically chosen class of accuracy from the model. For example, if an attacker specifically disallows nuclear facilities to be selectively recognized, it may be necessary to intentionally prevent unmanned aerial vehicles from correctly recognizing nuclear-related facilities. In this paper, we propose a selective poisoning attack that reduces the accuracy of only the chosen class in the model. The proposed method achieves this by training malicious data corresponding to only the chosen class while maintaining the accuracy of the remaining classes. For the experiment, we used tensorflow as the machine-learning library as well as MNIST, Fashion-MNIST, and CIFAR10 as the datasets. Experimental results show that the proposed method can reduce the accuracy of the chosen class by 43.2%, 41.7%, and 55.3% in MNIST, Fashion-MNIST, and CIFAR10, respectively, while maintaining the accuracy of the remaining classes. [ABSTRACT FROM AUTHOR]

In this paper, we describe a protocol framework that can perform classification tasks in a privacy-preserving manner. To demonstrate the feasibility of the proposed framework, we implement two protocols supporting Naive Bayes classification. We overcome the heavy computational load of conventional fully homomorphic encryption-based privacy-preserving protocols by using various optimization techniques. The proposed method differs from previous techniques insofar as it requires no intermediate interactions between the server and the client while executing the protocol, except for the mandatory interaction to obtain the decryption result of the encrypted classification output. As a result of this minimal interaction, the proposed method is relatively stable. Furthermore, the decryption key is used only once during the execution of the protocol, overcoming a potential security issue caused by the frequent exposure of the decryption key in memory. The proposed implementation uses a cryptographic primitive that is secure against attacks with quantum computers. Therefore, the framework described in this paper is expected to be robust against future quantum computer attacks. [ABSTRACT FROM AUTHOR]