Νευρωνικά δίκτυα και επεξεργασία φυσικής γλώσσας (Master thesis)

Σικαλόπουλος, Δημήτριος

Artificial Neural Networks (ANN) is a computational model inspired by Biological Neural Networks. ANNs have provided optimistic messages over recent years about research effectiveness and their reliable applications in Speech Recognition, Computer Vision, Text Processing, and so on. The purpose of this dissertation is to present the contribution of ANNs to their development and exploitation in the field of Natural Language Processing (NLP). Chapter 1 presents the important role of Grammar mathematical models in NLP and the most important stages in the evolution of Grammars that led to further methods and techniques. We develop the fundamental tools of ANNs, which are pioneering algorithms, as opposed to the traditional algorithms of the literature so far. These algorithms are defined on the one hand as Machine Learning Methods, that is, machine learning algorithms of the condition engine for dealing with specifications/data of a particular problem, and on the other hand as Deep Learning Methods, i.e. Machine Learning algorithms to deal with similar problem specifications. In addition, examples and comparisons of these with traditional algorithms are given. Chapter 2 is complemented by the presentation of the ANNs Methods, which is framed by definitions, Neural Network classifications, and examples of their applications. Chapter 3 lists the very recent methods and examples of the modern developments and applications of ANNs in NLP. Chapter 4 lists our suggestions for further research on ANNs and NLP. Specifically in the implementation of the Method of Modern Greek Language Processing proposed by: Template Grammars and Characteristic Exponents, Basic Modern Greek Language Computational Polylexicon and its algorithms. Finally, in the Appendices are included two glossaries of a Greek-English and an English-Greek basic dissertation terms correspondingly. At the beginning of each Chapter, keywords referred to the Chapter are listed and at the end of each Chapter, the corresponding bibliography is listed as well.
Institution and School/Department of submitter: Πανεπιστήμιο Ιωαννίνων. Σχολή Θετικών Επιστημών. Τμήμα Μαθηματικών
Subject classification: Νευρωνικά δίκτυα
Keywords: Νευρωνικά δίκτυα,Επεξεργασία φυσικής γλώσσας,Γραμματικές,Neural networks,Natural language processing,Grammars
URI: https://olympias.lib.uoi.gr/jspui/handle/123456789/29518
http://dx.doi.org/10.26268/heal.uoi.9593
Appears in Collections:Διατριβές Μεταπτυχιακής Έρευνας (Masters)

Files in This Item:
File Description SizeFormat 
Μ.Ε. ΣΙΚΑΛΟΠΟΥΛΟΣ ΔΗΜΗΤΡΙΟΣ 2019.pdf2.73 MBAdobe PDFView/Open


 Please use this identifier to cite or link to this item:
https://olympias.lib.uoi.gr/jspui/handle/123456789/29518
  This item is a favorite for 0 people.

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.