Search

This paper proposes a novel generative model called PUGAN, which progressively synthesizes high-quality audio in a raw waveform. Progressive upsampling GAN (PUGAN) leverages the progressive generation of higher-resolution output by stacking multiple encoder-decoder architectures. Compared to an existing state-of-the-art model called WaveGAN, which uses a single decoder architecture, our model generates audio signals and converts them to a higher resolution in a progressive manner, while using a significantly smaller number of parameters, e.g., 3.17x smaller for 16 kHz output, than WaveGAN. Our experiments show that the audio signals can be generated in real time with a comparable quality to that of WaveGAN in terms of the inception scores and human perception.

The 21st century, which we live in today, is rapidly changing based on the fourth industrial revolution and information and communication technology. According to change, existing urban models are emerging as smart cities. Smart City is a sustainable city that improves urban functions of various elements by applying information and communication convergence technology and environment friendly technology to urban space and solves urban problems that are occurring now. Energy, a core element of the city, enables efficient energy management of energy systems through an Internet of Things-based platform. The Internet of Things is a technology that sends and receives data by attaching sensors to various objects through a network. The goal is to use resources efficiently as the Internet of Things is spread and various industries are applied. This paper is a prediction study based on machine learning algorithms for efficient use of energy infrastructure in the Smart City. The proposed system builds the optimal Smart City by estimating energy efficiency of the Smart City and improving resource utilization.

Based on field induced atomic migration, one of the cycling endurance failure mechanisms in phase change memories, we propose an electrical treatment method to reduce atomic migration. The electric treatment is performed by applying compensation voltage after set or reset voltage applications, and we have a substantial enhancement of cycling endurance characteristics. The polarity of the compensation voltage is negative with respect to the programming voltages. To further investigate the effect of the electrical treatment, we introduced a parameter named melting voltage, which steadily increases as the number of set/reset cycling increases. This observation is consistent with the increase of resistance at the reset voltage, which supports void formation in the active region of Ge2Sb2Te5 with cycling stress.

IMT-Advanced system proposed by ITU-R for next generation mobile communications has the technical requirements of the performances of an improved broadcast, a multicast and a VoIP by supporting the transmission rates up to 100 Mbps for high mobility and 1 Gbps for low mobility. LTE-Advanced system is one of the leading candidates of IMT-Advanced systems and many telecommunications industries will introduce LTE-Advanced system widely. Therefore it is important to examine the performance of the technologies applied in LTE-Advanced system. In this paper, we analyze the performances of the transmission modes in the downlink spatial channel model of LTE-Advanced system.