Results 291 to 300 of about 725,384 (331)
Some of the next articles are maybe not open access.
Related searches:
Related searches:
Neural Processing Letters, 2021
This work proposes a new representation learning model called kernelized linear autoencoder. Instead of modeling non-linearity by the non-linear activation functions, we employ linear activations but account for non-linearity by the kernel trick. We propose four variants. The first one is the basic unsupervised kernelized linear autoencoder. The second
openaire +1 more source
This work proposes a new representation learning model called kernelized linear autoencoder. Instead of modeling non-linearity by the non-linear activation functions, we employ linear activations but account for non-linearity by the kernel trick. We propose four variants. The first one is the basic unsupervised kernelized linear autoencoder. The second
openaire +1 more source
Dimension of Kernels of Linear Operators
American Journal of Mathematics, 1992The basic question addressed in this paper is one of expressing the dimension of the intersection of kernels of linear operators that arise naturally in multivariate approximation theory in terms of the more easily computable dimensions of some basic blocks.
Jia, Rong-Qing +2 more
openaire +2 more sources
Acta Universitatis Nicolai Copernici, Ekonomia, 2012 The aim of this paper is to presents the results of simulation studies of size and power of the kernel linearity test, which belongs to a class of nonparametric tests. Simulation survey was carried out for linear and nonlinear models estimated by Ordinary Least Squares Method and Maximum Likelihood Method.
openaire
2009 Ninth IEEE International Conference on Data Mining, 2009
The design of a good kernel is fundamental for knowledge discovery from graph-structured data. Existing graph kernels exploit only limited information about the graph structures but are still computationally expensive. We propose a novel graph kernel based on the structural characteristics of graphs. The key is to represent node labels as binary arrays
Shohei Hido, Hisashi Kashima
openaire +1 more source
The design of a good kernel is fundamental for knowledge discovery from graph-structured data. Existing graph kernels exploit only limited information about the graph structures but are still computationally expensive. We propose a novel graph kernel based on the structural characteristics of graphs. The key is to represent node labels as binary arrays
Shohei Hido, Hisashi Kashima
openaire +1 more source
A quasi‐linear reproducing kernel particle method
International Journal for Numerical Methods in Engineering, 2016SummaryReproducing kernel particle method (RKPM) has been applied to many large deformation problems. RKPM relies on polynomial reproducing conditions to yield desired accuracy and convergence properties but requires appropriate kernel support coverage of neighboring nodes to ensure kernel stability.
Yreux, Edouard, Chen, Jiun-Shyan
openaire +2 more sources
Improved Kernel Learning Using Smoothing Parameter Based Linear Kernel
2003This article introduces a neural network capable of learning a temporal sequence. Directly inspired from a hippocampus model [2], this architecture allows an autonomous robot to learn how to imitate a sequence of movements with the correct timing. The results show that the network model is fast, accurate and robust.
A B M Shawkat Ali, Ajith Abraham
openaire +1 more source
Kernel Density Estimation on a Linear Network
Scandinavian Journal of Statistics, 2016AbstractThis paper develops a statistically principled approach to kernel density estimation on a network of lines, such as a road network. Existing heuristic techniques are reviewed, and their weaknesses are identified. The correct analogue of the Gaussian kernel is the ‘heat kernel’, the occupation density of Brownian motion on the network.
Mcswiggan, G. +2 more
openaire +2 more sources
Linear Replicator in Kernel Space
2010This paper presents a linear replicator [2][4] based on minimizing the reconstruction error [8][9] It can be used to study the learning behaviors of the kernel principal component analysis [10], the Hebbian algorithm for the principle component analysis (PCA) [8][9] and the iterative kernel PCA [3].
Wei-Chen Cheng, Cheng-Yuan Liou
openaire +1 more source