A multi-agent system to have information flow among agents is bound to experience a communication delay over a directed graph. However, delay information is often unavailable and time varying. To deal with this problem, a consensus control for a second-order multi-agent system consisting of a leader and multiple followers is developed. Sliding mode control is exploited to deal with uncertainties incurred by unknown time varying communication delay and disturbance. The proposed sliding mode consensus control is shown to achieve the asymptotic bounded consensus and the finite time convergence of a sliding variable for any unknown time varying delay with the bounded first and the second derivative. The numerical simulations verify the properties of the proposed algorithm and its performance which is almost identical to one achieved by a consensus algorithm exploiting known delay information.

With the evolution of the Internet of Things (IoT), various types of devices and massive systems comprising national infrastructures such as a smart grid are connected on a network, which poses various types of security issues in a cyber physical system. In this paper, we propose two false data injection attacks, which are on the forward path and the feedback path of a control system. Both are designed with a controllable parameter which determines the degree of degradation. A defensive method of inversing a linear forward attack through estimating with least square or minimum mean squared method was developed. A conventional Kalman filter was considered as a defensive method for a noise injection attack on the feedback path. The numerical evaluation verifies that the parameters of the proposed attacks control the degree of performance degradation of the control system, and the proposed defenses can effectively defend the proposed attacks.

We investigate the beamforming for cooperative non-orthogonal multiple access (NOMA) transmission through a decode-and-forward relay either with full-duplex (FD) mode or with half-duplex (HD) mode. The beamforming vectors at the source and at the relay are optimized to maximize the end-to-end sum rates of the NOMA transmission. The optimal beamforming for the second-hop of the NOMA transmission is present, from which the beamforming algorithms for the two relay modes are derived. For the FD case, the removal of self interference from the feedback loop is important and a two dimensional (2-D) search algorithm based on semi definite relaxation (SDR) is provided, where the 2-D search range is narrowed down as the algorithm iterates. For the HD mode, no iteration is necessary and only one implementation of SDR is required since the SI suppression is not needed. Simulation results show the performance advantage of the proposed FD mode system over that of HD mode in various parameter settings. This justifies the cost and complex hardware circuitry of FD mode.

The performance of existing face age progression or regression methods is often limited by the lack of sufficient data to train the model. To deal with this problem, we introduce a novel framework that exploits synthesized images to improve the performance. A conditional generative adversarial network (GAN) is first developed to generate facial images with targeted ages. The semi-supervised GAN, called SS-FaceGAN, is proposed. This approach considers synthesized images with a target age and the face images from the real data so that age and identity features can be explicitly utilized in the objective function of the network. We analyze the performance of our method over previous studies qualitatively and quantitatively. The experimental results show that the SS-FaceGAN model can produce realistic human faces in terms of both identity preservation and age preservation with the quantitative results of a decent face detection rate of 97% and similarity score of 0.30 on average.

Even though a certain degree of political bias is unavoidable in the media, strong media bias is likely to have an impact on society, especially on the formation of public opinion. This research proposes a data-driven method for quantifying political bias of media contents. With a document classification technique called doc2vec and social data from Facebook posts, a model for analysing the bias is developed. By applying the model to contents of major South Korean newspapers, this paper demonstrates quantitatively that significant political bias exists in the newspapers in line with the perceived political bias.

We propose signal-to-interference-plus-noise power ratio (SINR) maximizing rank one beamforming schemes for the full-duplex (FD) relay systems. The non-regenerative amplify and forward (AaF) relay is considered with multiple transmit antennas and multiple receive antennas. The residual selfinterference (SI) after analogue domain processing is to be suppressed by the proposed schemes so that the spectral efficiency of FD can be improved. We first analyze the correlation matrix of SI in a matrix power series and derive a convergent form of the correlation matrix so that the SINR expressions can be identified. Leveraging the structure of the SINR function on the beamformer (BF) vectors and power control parameter, two of the proposed schemes try to maximize SINR by optimizing a BF vector through iterative updates exploiting a bisection type search with the other BF vector being fixed. The third proposed scheme alternates over these two schemes to further enhance them by widening the search space that the BF vectors span. We compare the SINR of the proposed schemes with some reference schemes based on zero forcing approaches, maximum ratio reception, and maximum ratio transmission to characterize the performance improvement of the proposed schemes.