Although deep neural networks (DNNs) have achieved high performance across a variety of applications, they can often be deceived by adversarial examples that are generated by adding small perturbations to the original images. Adversaries may generate adversarial examples using the property of transferability, in which adversarial examples that deceive one model can also deceive other models because adversaries do not obtain any information on the DNNs deployed in real scenarios. Recent studies show that adversarial examples with feature space perturbations are more transferable than others. Adversarial training is an effective method to defend against adversarial attacks. However, it results in a decrease in the classification accuracy for natural images, and it is not sufficiently robust against transferable adversarial examples because it does not consider adversarial examples with feature space perturbations. We propose a novel adversarial training method to train DNNs to be robust against transferable adversarial examples and maximize their classification accuracy for natural images. The proposed method trains DNNs to correctly classify natural images and adversarial examples and also minimize the feature differences between them. The robustness of the proposed method was similar to those of the previous adversarial training methods for MNIST dataset and was up to average 6.13% and 9.24% more robust against transfer adversarial examples for CIFAR-10 and CIFAR-100 datasets, respectively. In addition, the proposed method yielded an average classification accuracy that was approximately 0.53%, 6.82%, and 10.60% greater than some state-of-the-art adversarial training methods for all datasets, respectively. The proposed method is robust against a variety of transferable adversarial examples, which enables its implementation in security applications that may benefit from high-performance classification but are at high risk of attack.

Strong designated verifier signature scheme (SDVS) allows a verifier to privately check the validity of a signature. Recently, Huang et al. first constructed an identity-based SDVS scheme (HYWS) in a stronger security model with non-interactive proof of knowledge, which holds the security properties of unforgeability, non-transferability, non-delegatability, and privacy of signer's identity. In this paper, we show that their scheme does not provide the claimed properties. Our analysis indicates that HYWS scheme neither resist on the designated verifier signature forgery nor provide simulation indistinguishability, which violates the security properties of unforgeability, non-delegatability and non-transferability.

In this paper, we propose an on-line e-coin system with four parties: Consumer, Merchant, Bank and Issuer. The proposed system not only circulates anonymous e-coins but also protects the profits of the Merchant and the Consumer during a transaction. An e-coin, consisting of a secret value c and a public value c'=h(c) where h() is a secure one-way hash function with collision resistant property, is generated by its owner. The public value of a legal e-coin is published on the bulletin board of Issuer. Only the owner who releases the secret values of the published e-coins can spend money. Instead of Bank, Issuer has to be on-line to verify and replace the public values of the Consumer's e-coins with the Merchant's while the Consumer pays money to the Merchant in a transaction. Such a replacement represents that the coins are passed from one person to another.