27 July 2020 in IEEE Transactions on Circuits and Systems II: Express Briefs
Deep neural networks are becoming increasingly popular. However, they are also vulnerable to adversarial attacks. The existing attack methods include white-box attack and black-box attack. The white-box attack assumes full model knowledge while the black-box one assumes none. In this brief, we propose a novel attack method between these two. Specifically, we have made the following contributions: (1) we propose the gray-box attack, which utilizes the side-channel attack to predict the model structure based on a pre-trained classifier and (2) we validate our method on real-world experiments. The experimental results show that our gray-box attack can significantly outperform the existing techniques.
31 March 2017 in International Journal of Systems Science
This paper investigates the problem of periodic motion control of second-order multi-agent systems in three dimensions under directed interaction topologies. Distributed algorithms for periodic rotating motion around a static and moving target were proposed by exploring the introduction of Cartestian coordinate coupling for each agent. In case of a static target, we show that when the nonsymmetric Laplacian matrix has certain properties, the damping gain is above a certain bound, and the Euler angle is equal to a critical value, all the agents will eventually rotate around the target periodically. In case of a moving target, the composite motion behaviours with translation and rotation will emerge, when this moving target's information can only be available to one or one subset of these agents and all agents have only local interactions. Tools like matrix theory, linear system theory and other mathematical skills were used for convergence analysis. Simulation results were provided to illustrate the effectiveness of the theoretical results.