This study deals with neural networks in the sense of geometric transformations acting on the coordinate representation of the underlying data manifold which the data is sampled from. It forms part of an attempt to construct a formalized general theory of neural networks in the setting of Riemannian geometry. From this perspective, the following theoretical results are developed and proven for ...

The nite di erence method is a numerical analysis technique used to solve problems involving irregular geometries, complicated boundary conditions, or both. The geometries are represented using partial di erential equations. The solutions to the partial di erential equations can be easily generated with the aid of a computer. As the geometries become increasingly complex, the solutions of the p...

This work is our rst attempt in establishing the connections between evolutionary computation algorithms and stochastic approximation procedures By treating evo lutionary algorithms as recursive stochastic procedures we study both constant gain and decreasing step size algorithms We formulate the problem in a rather general form supply the su cient conditions for convergence both with probabili...

image restoration has been an active research area. dierent formulations are eective in high qualityrecovery. partial dierential equations (pdes) have become an important tool in image processingand analysis. one of the earliest models based on pdes is perona-malik model that is a kindof anisotropic diusion (andi) lter. anisotropic diusion lter has become a valuable tool indierent elds...

generalized solution on neumann problem of the fourth order ordinary dierential equation in space w2 (0; b) has been discussed , we obtain the condition on b.v.p when the solution is in classical form. formulation of quintic spline function has been derived and the consistency relations are given.numerical method,based on quintic spline approximation has been developed. spline solution of the...

The e ciency of numerically solving time-dependent partial di erential equations on parallel computers can be greatly improved by computing the solution on many time-levels simultaneously. The theoretical properties of one such method, namely the discrete-time multigrid waveform relaxation method, are investigated for systems of ordinary di erential equations obtained by spatial nite element di...

in this paper, a new approach for solving the second order fuzzy differential equations (fde) with fuzzy initial value, under strongly generalized h-differentiability is presented. solving first order fuzzy differential equations by extending 1-cut solution of the original problem and solving fuzzy integro-differential equations has been investigated by some authors (see for example cite{darabi...

This paper considers the ®xed-structure, discrete-time mixed H2=H1 controller synthesis problem in the delta operator (di€ erence operator) framework. The di€ erential operator and shift operator versions of the problem are reviewed for comparison, and necessary conditions are derived for all three formulations. A quasi-Newton/continuation algorithm is then used to obtain approximate solutions ...

The paper introduces an approach for the derivation of discrete time approximations for solutions of stochastic di erential equations with time delay. The suggested approximations converge in a strong sense. Furthermore, explicit solutions for linear stochastic delay equations are given.

We describe the structure of the nonlinear system for the coe cients of diagonally implicit multistage integration methods for ordinary di erential equations. This structure is then utilized in the search for methods of high order and stage order with a given stability function. New methods were obtained by solving the nonlinear systems by state-of-the-art software based on least squares minimi...