41 1 0 0 OA 情報幾何学

著者
甘利 俊一
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.2, no.1, pp.37-56, 1992-03-16 (Released:2017-04-08)
参考文献数
26

Information geometry is a new theoretical method to elucidate intrinsic geometrical structures underlying information systems. It is applicable to wide areas of information sciences including statistics, information theory, systems theory, etc. More concretely, information geometry studies the intrinsic geometrical structure of the manifold of probability distributions. It is found that the manifold of probability distributions leads us to a new and rich differential geometrical theory. Since most of information sciences are closely related to probability distributions, it gives a powerful method to study their intrinsic structures. A manifold consisting of a smooth family of probability distributions has a unique invariant Riemannian metric given by the Fisher information. It admits a one-parameter family of invariant affine connections, called the α-connection, where α and-α-connections are dually coupled with the Riemannian metric. The duality in affine connections is a new concept in differential geometry. When a manifold is dually flat, it admits an invariant divergence measure for which a generalized Pythagorian theorem and a projection theorem hold. The dual structure of such manifolds can be applied to statistical inference, multiterminal information theory, control systems theory, neural networks manifolds, etc. It has potential ability to be applied to general disciplines including physical and engineering sciences.
著者
佐藤 寛之
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.27, no.1, pp.21-30, 2017 (Released:2017-06-30)
参考文献数
29

This paper deals with Riemannian optimization, that is, optimization on Riemannian manifolds. Theories of Euclidean optimization and Riemannian manifolds are first briefly reviewed together with some simple and motivating examples, followed by the Riemannian optimization theory. Retractions and vector transports on Riemannian manifolds are introduced according to the literature to describe a general Riemannian optimization algorithm. Recent convergence analysis results of several types of Riemannian conjugate gradient methods, such as Fletcher-Reeves and Dai-Yuan-types, are then given and discussed in detail. Some applications of Riemannian optimization to problems of current interest, such as 1)singular value decomposition in numerical linear algebra; 2)canonical correlation analysis and topographic independent component analysis as statistical methods; 3)low-rank tensor completion for machine learning; 4)optimal model reduction in control theory; and 5)doubly stochastic inverse eigenvalue problem, are also introduced.
著者
石井 晃 太田 奨
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.25, no.2, pp.50-58, 2015-06-25 (Released:2017-04-08)

We apply a mathematical theory for hit phenomenon for prediction of the "general election" of AKB48 which is very popular girls group in Japan.
著者
斉藤 一哉
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.26, no.1, pp.9-14, 2016 (Released:2016-07-27)
参考文献数
18

This study presents a new method for designing self-deploying origami using the geometrically misaligned creases. In this method, some facets are replaced by “holes” such that the systems become a 1-DOF mechanism. These perforated origami models can be folded and unfolded similar to rigidfoldable(without misalignment) models because of their DOF despite the existence of the misalignment. Focusing on the removed facets, the holes will deform according to the motion of the frame of the remaining parts. In the proposed method, these holes are filled with elastic parts and store elastic energy for self-deployment. First, a new extended rigid-folding simulation technique is proposed to estimate the deformation of the holes. Next by using the above technique, the proposed method is applied on arbitrary-size quadrilateral mesh origami. Finally, by using the finite-element method, the authors conduct numerical simulations and confirm the deployment capabilities of the models.
著者
今野 浩
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.1, no.1, pp.36-50, 1991-03-15 (Released:2017-04-08)

Recently, a remarkable progress has been made in the field of the global minimization of nonconvex functions over a polytope. The purpose of this article is to survey one of the most successful approaches in this field, namely parametric programming approaches to quasilinear nonconvex minimization problems. The problems to be discussed are: linear multiplicative programming problems, i. e., the minimization of the product of two affine functions; minimization of the sum of two linear fractional functions; minimization of concave quadratic functions and bilinear programming problems. It will be shown that a global minimum of a fairly large scale problems can be obtained efficiently by applying parametric simplex algorithms. Further, it will be shown that a convex multiplicative programming problems, i. e., the minimization of the product of two convex functions, can be solved by parametrizatioh and branch and bound techniques.
著者
青木 康憲 速水 謙 小長谷 明彦
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.24, no.4, pp.151-159, 2014-12-25 (Released:2017-04-08)

As the observations we can make from patients are limited compared to the complexity of the physiology, underdetermined inverse problems appear often in the parameter estimation problems of physiologically based pharmacokinetics (PBPK) models. We address this issues of not being able to identify the model parameter set uniquely by finding multiple sets of possible parameter sets that are consistent with the observations. As this approach requires multiple parameter estimations of a complex model, the computational cost can be a bottle neck. In this paper, we introduce a new computationally efficient algorithm called the Cluster Newton method to find multiple solutions of an underdetermined inverse problem.
著者
牧野 淳一郎
出版者
一般社団法人 日本応用数理学会
雑誌
応用数理 (ISSN:24321982)
巻号頁・発行日
vol.8, no.4, pp.277-287, 1998-12-15 (Released:2017-04-08)
参考文献数
18

I overview the Fast Multipole Method (FMM) and the Barnes-Hut tree method. These algorithms evaluate mutual gravitational interaction between N particles in O(N) or O(N log N) times, respectively. I present basic algorithms as well as recent developments, such as Anderson's method of using Poisson's formula, the use of FFT, and other optimization techniques. I also summarize the current states of two algorithms. Though FMM with O(N) scaling is theoretically preferred over O(N log N) tree method, comparisons of existing implementations proved otherwise. This result is not surprizing, since the calculation cost of FMM scales as O(Np^2) where p is the order of expansion, while that of the tree method scales as O(N log Np).