We assume that neurons have a strongly non-degenerate property in the mean including differentiation, express 3-layered neural networks as neurovarieties in function space with an inner product, and define a gradient flow of error-functions by the inner product. We study the gradient flow in the regular points area of the neurovariety. As a result, it is proved that Hessian matrices at the fixed points are non-degenerate and the fixed points are isolated, for almost every target function which is not on the neurovariety. Furthermore we can say that long (serious) plateaus arise only in the singular points area of the neurovariety.
The House of Representatives has 480 members and 300 members out of them are apportioned among 47 prefectures. Each apportionment is evaluated generally with respect to the criterion: the ratio of the maximum district size to the minimum district size. This paper shows that this criterion does not give an apportionment proportional to the population of prefectures. The actual apportionment is determined by the method of apportioning 253 members among 47 prefectures under the method of greatest remainders and adding one member to each prefecture. In addition, it is shown that this method is not proportional either.
A variational principle that minimizes the errors and noises associated with experimental measurements of two-dimensional non-steady-state temperature fields is investigated. The variational principle also assures the satisfaction of the governing time-developing heat conduction equation. On the basis of the variational principle, an intelligent hybrid experimental-numerical method is formulated and a scheme of calculation is presented. Furthermore, a numerical simulation is illustrated. The present paper is the extension of studies in steady-state temperature fields and is the development of researches in elastic deformation fields, elastic-plastic deformation fields, and the near-field of an interface crack tip.
We introduce extended autonomous Hamiltonian systems which are equivalent to non-autonomous Hamiltonian systems. Applying the high-order energy conserving integrators to the extended systems, we construct integrators for the non-autonomous systems, by which we can calculate the changes of energy with an accuracy of high-order.