The fundamentals and developments of information processing hardware systems using high-dimensional complex dynamics are explained. In particular, how I conceived the idea of the information processing paradigm through nonlinear complex hardware systems and how I implemented the real-number computation with noise in analog integrated circuit systems are described together with my personal experiences. In addition, it is shown that the constructive approach, in which complexity is handled as complex as it is, is effective for the analysis and design of nonlinear complex engineering systems, where both reductionism and holism would fail, which are usually employed for linear and digital computing systems.
Quantum computers are expected to have an overwhelming computational power that the current computer technology cannot achieve. To derive such a computational power from quantum hardware, quantum algorithms (i.e., algorithms specialized to quantum computers) are crucial. In this paper, the quantum algorithms for a single quantum computer and those for multiple quantum computers that can communicate with each other are reviewed, and then, basic ideas underlying the quantum algorithms that the current authors have developed are illustrated.
In the fall of the last year, we finished the Japanese translation of the already classical masterpiece, “Winning Ways for Mathematical Plays” by E.R. Berlekamp, J.H. Conway and R.K. Guy into Japanese. On the occasion of SITA (Symposium of Information Theory and its Applications) 2019, Kirishima, I gave a lecture on Combinatorial Game Theory tributing Prof. Berlekamp who passed away on 9 April 2019, for his great achievements on Information Theory and Combinatorial Game Theory. Here I give an overview of CGT and related topics presented at the symposium. Combinatorial games are two-person games without any chance elements and with no hidden information. They include child's play such as Tic-Tac-Toe, and more deeper board games such as Go, Chess and Shogi. For your good plays, you need to know the values of game positions that are considered as the extended concept of numbers. We can observe the new idea on evaluating the positions such as the surreal numbers invented by Prof. Conway.
Flying drones (multirotors) are expected to bring tremendous benefit to our industry and society. However, there is a potential risk to civil safety if multirotors crash. Therefore, it is important to establish multirotor flight controls in the case of complete motor failures. In this paper we discuss recent research activities on multirotor motor failure problems, focusing on the problems of multirotor flight states to avoid a crash in the case of complete motor failure. In addition, by the Euler angle state variable approach, research results on the relationship the between operating points (equilibrium points) of multirotors and the flight states to avoid a crash in the case of complete motor failure, are also reviewed.
Deep learning techniques can be used for not only learning deep neural networks but also internal parameter optimization of differentiable iterative algorithms. By embedding learnable parameters into an excellent iterative algorithm, we can construct a flexible derived algorithm with data-driven learnability. This approach is called deep unfolding. We present an overview of deep unfolding and its features, focusing on sparse signal recovery algorithms. In the first half of this paper, examples of deep unfolding including a sparse signal recovery algorithm, TISTA, will be presented. We observed the convergence acceleration phenomenon for deep unfolding-based algorithms. In the second half of the paper, our theoretical results (spectral radius control based on the Chebyshev step) for convergence acceleration are outlined.