Search results for "approximation"

Experiments in Value Function Approximation with Sparse Support Vector Regression

2004

We present first experiments using Support Vector Regression as function approximator for an on-line, sarsa-like reinforcement learner. To overcome the batch nature of SVR two ideas are employed. The first is sparse greedy approximation: the data is projected onto the subspace spanned by only a small subset of the original data (in feature space). This subset can be built up in an on-line fashion. Second, we use the sparsified data to solve a reduced quadratic problem, where the number of variables is independent of the total number of training samples seen. The feasability of this approach is demonstrated on two common toy-problems.

Multi-dimensional Function Approximation and Regression Estimation

2002

In this communication, we generalize the Support Vector Machines (SVM) for regression estimation and function approximation to multi-dimensional problems. We propose a multi-dimensional Support Vector Regressor (MSVR) that uses a cost function with a hyperspherical insensitive zone, capable of obtaining better predictions than using an SVM independently for each dimension. The resolution of the MSVR is achieved by an iterative procedure over the Karush-Kuhn-Tucker conditions. The proposed algorithm is illustrated by computers experiments.

Space partitioning of exchange-correlation functionals with the projector augmented-wave method

2018

We implement a Becke fuzzy cells type space partitioning scheme for the purposes of exchange-correlation within the GPAW projector augmented-wave method based density functional theory code. Space partitioning is needed in the situation where one needs to treat different parts of a combined system with different exchange-correlation functionals. For example, bulk and surface regions of a system could be treated with functionals that are specifically designed to capture the distinct physics of those regions. Here, we use the space partitioning scheme to implement the quasi-nonuniform exchange-correlation scheme, which is a useful practical approach for calculating metallic alloys on the gene…

Appendix: Diophantine Approximation on Hyperbolic Surfaces

2002

In this (independent) appendix, we study the Diophantine approximation properties for the particular case of the cusped hyperbolic surfaces, in the spirit of Sect. 2 (or [11]), and the many still open questions that arise for them. We refer to [9], [10]for fundamental results and further developments. We study in particular the distance to a cusp of closed geodesics on a hyperbolic surface.

Multiresolution Analysis for Irregular Meshes

2003

International audience; The concept of multiresolution analysis applied to irregular meshes has become more and more important. Previous contributions proposed a variety of methods using simplification and/or subdivision algorithms to build a mesh pyramid. In this paper, we propose a multiresolution analysis framework for irregular meshes with attributes. Our framework is based on simplification and subdivision algorithms to build a mesh pyramid. We introduce a surface relaxation operator that allows to build a non-uniform subdivision for a low computational cost. Furthermore, we generalize the relaxationoperator to attributes such as color, texture, temperature, etc. The attribute analysis…

Closed Form Approximation of Swap Exposures

2013

This paper provides closed form lower and upper bounds for the price of European swaption on cross currency basis swap with the presence of dynamic basis spreads. Cross currency basis spreads are treated as integrals of spot spreads, approach familiar from interest rate models. The spot spread is modelled by two-factor mean reverting Gaussian model that is equivalent to two-factor Hull-White model introduced by [Hull and White(1994)]. This model allows closed form approximations and relatively well fitting and simple calibration to the spread term structure.

Discontinuous Galerkin semi-Lagrangian method for Vlasov-Poisson

2011

We present a discontinuous Galerkin scheme for the numerical approximation of the one-dimensional periodic Vlasov-Poisson equation. The scheme is based on a Galerkin-characteristics method in which the distribution function is projected onto a space of discontinuous functions. We present comparisons with a semi-Lagrangian method to emphasize the good behavior of this scheme when applied to Vlasov-Poisson test cases. Une méthode de Galerkin discontinu est proposée pour l’approximation numérique de l’équation de Vlasov-Poisson 1D. L’approche est basée sur une méthode Galerkin-caractéristiques où la fonction de distribution est projetée sur un espace de fonctions discontinues. En particulier, …

Rectifiability, weak linear approximation and tangent measures

1995

Théorie des spectres rovibroniques des molécules octaédriques : Hamiltonien et moments de transition

2002

This thesis is devoted to the treatment of rovibronic couplings of octahedral species for which the Born-Oppenheimer approximation is broken down. By using the octahedral formalism, a full effective rovibronic model is extended from works about molecules in a non-degenerate electronic state. This effective model is dedicated to molecules with an odd or an even number of electrons and it has been successfully applied to V(CO)6 and ReF6. For both of them we have four interacting vibronic sublevels attributed to a dynamical Jahn-Teller effect and giving rise to very complicated spectra. This model is validated by the overall agreement between predicted and observed band profiles. Moreover, an …

Non-Linear Hysteretic Instability in Rotating Machinery

2013