Search results for "kernel"

Learning the relevant image features with multiple kernels

2009

This paper proposes to learn the relevant features of remote sensing images for automatic spatio-spectral classification with the automatic optimization of multiple kernels. The method consists of building dedicated kernels for different sets of bands, contextual or textural features. The optimal linear combination of kernels is optimized through gradient descent on the support vector machine (SVM) objective function. Since a na¨ive implementation is computationally demanding, we propose an efficient model selection procedure based on kernel alignment. The result is a weight — learned from the data — for each kernel where both relevant and meaningless image features emerge after training. E…

An efficient method for clustered multi-metric learning

2019

Abstract Distance metric learning, which aims at finding a distance metric that separates examples of one class from examples of the other classes, is the key to the success of many machine learning tasks. Although there has been an increasing interest in this field, learning a global distance metric is insufficient to obtain satisfactory results when dealing with heterogeneously distributed data. A simple solution to tackle this kind of data is based on kernel embedding methods. However, it quickly becomes computationally intractable as the number of examples increases. In this paper, we propose an efficient method that learns multiple local distance metrics instead of a single global one.…

Mapping properties of weakly singular periodic volume potentials in Roumieu classes

2020

The analysis of the dependence of integral operators on perturbations plays an important role in the study of inverse problems and of perturbed boundary value problems. In this paper, we focus on the mapping properties of the volume potentials with weakly singular periodic kernels. Our main result is to prove that the map which takes a density function and a periodic kernel to a (suitable restriction of the) volume potential is bilinear and continuous with values in a Roumieu class of analytic functions. This result extends to the periodic case of some previous results obtained by the authors for nonperiodic potentials, and it is motivated by the study of perturbation problems for the solut…

The Poisson problem: A comparison between two approaches based on SPH method

2012

Abstract In this paper two approaches to solve the Poisson problem are presented and compared. The computational schemes are based on Smoothed Particle Hydrodynamics method which is able to perform an integral representation by means of a smoothing kernel function by involving domain particles in the discrete formulation. The first approach is derived by means of the variational formulation of the Poisson problem, while the second one is a direct differential method. Numerical examples on different domain geometries are implemented to verify and compare the proposed approaches; the computational efficiency of the developed methods is also studied.

The smoothed particle hydrodynamics method via residual iteration

2019

Abstract In this paper we propose for the first time an iterative approach of the Smoothed Particle Hydrodynamics (SPH) method. The method is widespread in many areas of science and engineering and despite its extensive application it suffers from several drawbacks due to inaccurate approximation at boundaries and at irregular interior regions. The presented iterative process improves the accuracy of the standard method by updating the initial estimates iterating on the residuals. It is appealing preserving the matrix-free nature of the method and avoiding to modify the kernel function . Moreover the process refines the SPH estimates and it is not affected by disordered data distribution. W…

The First Main Theorem

1998

Operational cloud screening service for Sentinel-2 image time series

2015

This paper deals with the development and implementation of a cloud screening algorithm for image time series, with the focus on the forthcoming Sentinel-2 satellites to be launched under the ESA Copernicus Programme. The proposed methodology is based on kernel ridge regression and exploits the temporal information to detect anomalous changes that correspond to cloud covers. The huge data volumes to be processed when dealing with high temporal, spatial, and spectral resolution datasets motivate the implementation of the algorithm within distributed computer resources. In consequence, an operational cloud screening service has been specifically designed and implemented in the frame of the Se…

Fast Approximated Discriminative Common Vectors Using Rank-One SVD Updates

2013

An efficient incremental approach to the discriminative common vector (DCV) method for dimensionality reduction and classification is presented. The proposal consists of a rank-one update along with an adaptive restriction on the rank of the null space which leads to an approximate but convenient solution. The algorithm can be implemented very efficiently in terms of matrix operations and space complexity, which enables its use in large-scale dynamic application domains. Deep comparative experimentation using publicly available high dimensional image datasets has been carried out in order to properly assess the proposed algorithm against several recent incremental formulations.

Heterogeneous PBLAS: Optimization of PBLAS for Heterogeneous Computational Clusters

2008

This paper presents a package, called Heterogeneous PBLAS (HeteroPBLAS), which is built on top of PBLAS and provides optimized parallel basic linear algebra subprograms for heterogeneous computational clusters. We present the user interface and the software hierarchy of the first research implementation of HeteroPBLAS. This is the first step towards the development of a parallel linear algebra package for heterogeneous computational clusters. We demonstrate the efficiency of the HeteroPBLAS programs on a homogeneous computing cluster and a heterogeneous computing cluster.

Kernel estimation and display of a five-dimensional conditional intensity function

2018

The aim of this paper is to find a convenient and effective method of displaying some second order properties in a neighbourhood of a selected point of the process. The used techniques are based on very general high-dimensional nonparametric smoothing developed to define a more gen- eral version of the conditional intensity function introduced in earlier earthquake studies by Vere-Jones (1978). 1976) is commonly used for such a purpose in discussing the cumulative behavior of interpoint distances about an initial point. It is defined as the expected number of events falling within a given distance of the initial event, divided by the overall density (rate in 2-dimensions) of the process, sa…