New hybrid kernel architectures for deep learning

View/Open
Document typeMaster thesis
Date2018-04
Rights accessOpen Access
All rights reserved. This work is protected by the corresponding intellectual and industrial
property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public
communication or transformation of this work are prohibited without permission of the copyright holder
Abstract
In this work we explore the possibilities of combining neural network architectures and kernel methods by introducing hybrid kernel blocks. We present hybrid architectures which can be trained as traditional neural networks and introduce novel training and regularization methodologies for them.
SubjectsNeural networks (Computer science), Machine learning, Kernel functions, Xarxes neuronals (Informàtica), Aprenentatge automàtic, Kernel, Funcions de
DegreeMÀSTER UNIVERSITARI EN INTEL·LIGÈNCIA ARTIFICIAL (Pla 2012)
Collections
Files | Description | Size | Format | View |
---|---|---|---|---|
133133.pdf | 7,221Mb | View/Open |