Yuya Dan

Mathematical Analysis and Simulation of Information Diffusion on Networks

It is essential to consider the structure of information media in order to investigate the phenomena of information diffusion. As an example of continuous media, we consider on Euclidean spaces in the ordinary geometry. We also know information networks which have a variety of metric and structure in the space as another example of discrete media. In this work, we are faced the limit of mathematical analysis of nonlinear structure of information networks, then perform the Monte Carlo computer simulation using the proposed method.

Information Diffusion and Dissipative Effect on Social Networks

It is quite important in ITeS to consider phenomena and behavior of people on social networks. In this chapter, we discuss information diffusion and dissipative effect using analytical and computational methods in order to investigate the phenomena and behavior. In general, network structure plays an essential role in the process of information diffusion so that we are faced by the difficulties in control our personal and confidential information. The computer simulation in this chapter assumes a scale-free network as often seen in online social networks. We find out that the process of information diffusion obeys a certain curve and that there may be quantum leaps which are caused by the hubs of social networks. Adding to this, we have obtained the results of dissipative effect using two operations. One eliminates the main hub from the networks, the other reconstructs the network as a stochastic network based on a constant diffusion probability. Moreover, we obtain the suggestion of information diffusion and dissipative effect for the further investigation in information management and malware infection in information security.

Grid Computing for Artificial Intelligence

Extensive deployment of AI services, especially mobile AI, requires the support of edge computing. This support is not just at the network architecture level, the design, adaptation, and optimization of edge hardware and software are equally important. Specifically, (1) customized edge hardware and corresponding optimized software frameworks and libraries can help AI execution more efficiently; (2) the edge computing architecture can enable the offloading of AI computation; (3) well-designed edge computing frameworks can better maintain AI services running on the edge; (4) fair platforms for evaluating Edge AI performance help further evolve the above implementations.

Possibility of Human Grid Computing for Artificial Intelligence Systems

The paper is concerned in human grid computing for artificial intelligence systems. In general, grid computing enables to process a huge amounts of data in any fields, however, grid systems for Shogi, or Japanese chess, need to process astronomical combinations of positions. It is described that the proposed system is effective in reduction of computational complexity with collaboration from human intuition. A solution to the next move problem as artificial intelligence systems will be discussed. The paper introduces the current state of computer Shogi systems, and then proposes an idea of the interaction between human and grid computing in order to solve the next move problem in game theory.