A statistical manifold associated with a correlated walk (CW) model is examined by noticing a non-Riemannian curvature, called the α-curvature, as well as the Riemann curvature. Dynamical characteristics of the Riemann curvature and the α-curvature are discussed to have a close relation to the stability of the CW system. The statistical manifold is also found to be asymptotically flat in the meaning of the α (=1)-curvature, and also the jump probabilities characterizing the CW model is shown to have relation to a symmetry of the statistical manifold. Moreover a forecast is given about the statistical manifolds of n-step correlated-walk models and nonlinear models, and also such time-developing statistical manifolds are shown to be very analogous to the geometrical structure of Newton-Cartan theory of gravity.

View full abstract

Let G be a Lie group. A TC-pair (ρ,M ) of G gives a twisted linear action on a sphere. Assume that G ⁄kerρ is non-compact. Then the TC-pairs are equivalent, if the corresponding twisted linear actions are coincides with each other. Hence also so, if the corresponding twisted linear actions are transitive and equivariantly diffeomorphic.

View full abstract

Let H be a separable Hilbert space and let A be the algebra of continuous functions on the torus T ² which are uniform limits of polynomials in e ^imxe ^iny where

(m,n)∈{(m,0)∈Z ²|m ≥ 0}∪{(m,n)∈Z ²|n ≥ 1}.

For this uniform algebra A, we characterize invariant subspaces of L_H².

View full abstract

A variation of the Jentzsch theorem is given; the largest characteristic value with respect to odd characteristic functions is simple for an integral equation whose kernel is non-negative, symmetric, even on one of variables in R ⁿ and having some other conditions. In order to prove the new theorem, we present a simple proof for the Jentzsch theorem for a non-negative and symmetric integral kernel. We give an application of the new theorem to correlation functions of the one-dimensional spin systems.

View full abstract

A positive reinforcement type learning algorithm is formulated for a stochastic feed-forward multilayer neural network, with far interlayer synaptic connections, and we obtain a learning rule similar to that of the Boltzmann machine on the same multilayer structure. By applying a mean field approximation to the stochastic feed-forward neural network, the generalized error back-propagation learning rule is derived for a deterministic analog feed-forward multilayer network with the far interlayer synaptic connections.

View full abstract

In this paper, we study the orders of the two errors, phase and amplification errors, of explicit parallel Runge-Kutta methods for the integrations of periodic initial value problems. We first give a mathematical definition of the effective stages of the parallel Runge-Kutta method. Using the result, we show that the orders of these two errors depend on the number of the effective stages in parallel executions of the methods, and that the sum of these orders is bounded by 2s +1, where s is the number of the effective stages.

View full abstract

In spite of the fact that subspace method can approximate the distribution of categories precisely, only a few attempts have so far been made at applying it in hand-written character recognition. The subspace method proposed by Watanabe (1969) offers the basic concept of subspace construction, but the issue of how to use the limited samples to construct effective subspace to avoid the problem of mis-recognition caused by the fact that the subspace of a category is almost parallel to the mean vector of the category remains unresolved. This problem leads to the mis-recognition of the samples far from the mean vector. To cope with this problem, Abe, Nemoto and Sun (1995) have proposed the Combination method (CM), which constructs the subspace from several groups including different number of samples divided from the whole training samples. CM obtained a high recognition rate of 97.76% with respect to ETL9B, the largest database of hand-written characters in Japan. ETL9B was published by Electro-technical Laboratory of Japan in 1985. It includes 2,965 categories of Chinese characters and 71 categories of Japanese Kana, there are 200 data sets in the database where each set has one sample per category. Total number of samples is 607,200. The issues that need to be dealt with next are how to improve the recognition accuracy and how to accelerate the recognition speed. In this paper, we propose a new method called Uniform Division Method (UDM), which uses the uniformly divided training samples to construct subspace. Compared to CM given earlier, UDM is very simple and effective enough to improve the accuracy of recognition; as a result, we obtained a recognition rate of 98.64% for ETL9B compared to the 97.76% for CM. This is the first time that such a high recognition rate has been obtained by making good use of subspace method. Furthermore, the computation required for UDM is less than a half of that of CM. The UDM algorithm and the experiments with ETL9B will be described in this paper.

View full abstract

The cellular concentration of Paramecium protein kinase A (PKA) exhibits a bi-modal pattern in a 12-h light and 12-h dark cycle, with peaks around midday and immediately after dusk, in contrast to that measured in constant darkness (DD) which exhibits a mono-phasic pattern with a trough after subjective midday and a peak around subjective midnight. The bi-modal increases in PKA concentration caused by the onset (dawn) and offset (dusk) of illumination might contribute discrete phase-advances of the motility rhythm in P. multimicronucleatum. Fluctuation of PKA concentrations in DD is significantly related to that of frequency of changing moving-direction, on which PKA concentrations elevated during illumination may be additively observed to cause the bi-modal pattern.

View full abstract

The explosion in the amount of electronically available information and the powerful computers at lower cost are causing the computer-for-everybody trend, extending the range of users to non-experts. These phenomena create an information superhighway, which would be difficult to approach without help. Intelligent software agents will provide this necessary help to navigate in the world-wide network of information. In this paper, we survey the current status of agent-based concepts and applications. In particular, we especially review the agent-based information locating systems and present our view of the trends of agent technologies supporting users in finding information on distributed heterogeneous environments.

View full abstract

This paper provides with a new framework for traffic network equilibria with incomplete information. The basic element of our approach is differential information; different drivers have different information about their environment; they choose routes based on their private (differentiated) information. The purpose of this paper is to develop a general equilibrium model that makes explicit information or beliefs that a driver has as part of his/her primitive characteristics. The model we present is a reinterpretation on Harsanyi’s incomplete information game in a network game context. The difference from Harsanyi’s approach is the explicit consideration of rational expectations formation by drivers. A numerical illustration may provide us with a pedagogical insight on traffic network equilibria with incomplete information.

View full abstract

This paper is concerned with the social construction of the body in information society. In a word, how is the body viewed and treated by postmodern society via its vehicles of communication. To answer this question we focus on television commercials in two information-based countries: America and Japan. The rationale for medium selection and comparative focus are specified in the text.
The research involved the systematic collection of over 7,000 ads from the 7 major networks (4 in Japan, 3 in America) in cities of roughly comparable geo-political and demographic characteristics. Via theoretical sampling, a subset of 2353 (1221 from Japan, 1132 from America) was selected. Coding revealed that 581 (or 25%) involved the body in non-trivial ways. The presentation and use of body differed by country: 30% of American advertising versus 20% of Japanese CMs contained body content.
Four working hypotheses guided analysis. H1: two vastly different societies would commercialize the body in ways more similar than different; H2: a high degree of “body content” would exist in both societies; H3: “body consciousness”—i.e. intentional use of the body—would be higher in America than in Japan; H4: recurrent tendencies in body presentation would be uncovered. With specifiable qualifications, all four hypotheses were confirmed.
Four key findings emerged from this research. (1) confirmation of converging culture; (2) the narrow, repetitious and powerful messages in commercial media; (3) the centrality of societal images concerning identity and personal behavior (as communicated via the body); and (4) the emerging importance of the body as a social object in America and Japan. I conclude that the commercialized body is an increasingly central object in contemporary society.

View full abstract

This article presents an analysis of what is called VP Deletion (VPD) in English. Of particular concern is how the behavioral difference of VP-adverbs in a deletion construction can be accounted for and how the target of the rule operation can be specified. Pointing out that there is a syntactic distributional variation among VP-adverbs and this is what is responsible for the deletability difference in a VPD construction, we argue that the operation range of VPD can be fixed on a particular projection. It is also claimed that the present analysis contributes to making uncecessary Sag(1976)’s filter that requires the deletion site to follow AUX.

View full abstract