Many works have shown that strong connections relate learning from examples to regularization techniques for ill-posed inverse problems. Nevertheless by now there was no formal evidence neither that learning from examples could be seen as an inverse problem nor that theoretical results in learning theory could be independently derived using tools from regularization theory. In this paper we pro...

Point neuron models with a Heaviside firing rate function can be ill-posed. That is, the initial-condition-to-solution map might become discontinuous in finite time. If a Lipschitz continuous but steep firing rate function is employed, then standard ODE theory implies that such models are well-posed and can thus, approximately, be solved with finite precision arithmetic. We investigate whether ...

GMRES is one of the most popular iterative methods for the solution of large linear systems of equations. However, GMRES generally does not perform well when applied to the solution of linear systems of equations that arise from the discretization of linear ill-posed problems with error-contaminated data represented by the right-hand side. Such linear systems are commonly referred to as linear ...

Dynamically re-programmable hardware (e.g. Field Programmable Gate Arrays or FPGAs) change many of our basic assumptions of what hardware is. Normal hardware design focuses on the development of a static circuit of xed size, topology and functionality. The static nature of the target design entity is re ected in the design of CAD tools, hardware description languages, net-list representations a...

We consider large scale ill-conditioned linear systems arising from discretization of ill-posed problems. Regularization is imposed through an (assumed known) upper bound constraint on the solution. An iterative scheme, requiring the computation of the smallest eigenvalue and corresponding eigenvector, is used to determine the proper level of regularization. In this paper we consider several co...

Convergent methodology for ill-posed problems is typically equivalent to application of an operator dependent on a single parameter derived from the noise level and the data (a regularization parameter or terminal iteration number). In the context of a given problem discretized for purposes of numerical analysis, these methods can be viewed as resulting from imposed prior constraints bearing th...

We give an effective sufficient condition for a variational problem with infinite horizon on a Riemannian manifold M to admit a smooth optimal synthesis, i. e. a smooth dynamical system on M whose positive semi-trajectories are solutions to the problem. To realize the synthesis we construct a well-projected to M invariant Lagrange submanifold of the extremals’ flow in the cotangent bundle T ∗M ...

Multilevel methods are popular for the solution of well-posed problems, such as certain boundary value problems for partial differential equations and Fredholm integral equations of the second kind. However, little is known about the behavior of multilevel methods when applied to the solution of linear ill-posed problems, such as Fredholm integral equations of the first kind, with a right-hand ...

Nonlinear diffusions have successfully been employed to perform a variety of tasks in image processing. In the context of denoising, the Perona-Malik equation is a particularly successful example. In spite of its cherished qualities, it is mathematically ill-posed and has some practical short-comings. In this short paper a new nonlocal nonlinear diffusion is reported upon, which is locally well...

In this paper we solve a stabilization and tracking problem for linear systems, using a low gain controller suggested by the internal model principle. These results are a partial extension of results by Hämäläinen and Pohjolainen. In their results the plant is required to have a transfer function in the Callier-Desoer algebra, while in this paper we only require the system to be well-posed. The...