prikaz prve stranice dokumenta Razumijevanje prirodnog jezika korištenjem dubokog učenja
  Download
PDF 2.27 MB
undergraduate thesis
Razumijevanje prirodnog jezika korištenjem dubokog učenja
2021. urn:nbn:hr:200:016691
Müler, Marko
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Software Engineering
Chair of Visual Computing

Cite this document

Müler, M. (2021). Razumijevanje prirodnog jezika korištenjem dubokog učenja (Undergraduate thesis). Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek. Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:016691

Müler, Marko. "Razumijevanje prirodnog jezika korištenjem dubokog učenja." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2021. https://urn.nsk.hr/urn:nbn:hr:200:016691

Müler, Marko. "Razumijevanje prirodnog jezika korištenjem dubokog učenja." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2021. https://urn.nsk.hr/urn:nbn:hr:200:016691

Müler, M. (2021). 'Razumijevanje prirodnog jezika korištenjem dubokog učenja', Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 25 April 2024, https://urn.nsk.hr/urn:nbn:hr:200:016691

Müler M. Razumijevanje prirodnog jezika korištenjem dubokog učenja [Undergraduate thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2021 [cited 2024 April 25] Available at: https://urn.nsk.hr/urn:nbn:hr:200:016691

M. Müler, "Razumijevanje prirodnog jezika korištenjem dubokog učenja", Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2021. Available at: https://urn.nsk.hr/urn:nbn:hr:200:016691

Please login to the repository to save this object to your list.
Metadata
TitleRazumijevanje prirodnog jezika korištenjem dubokog učenja
Title (english)Understanding natural language using deep learning
AuthorMarko Müler
MentorIrena Galić (mentor)
MentorHrvoje Leventić (sumentor)
Committee memberIrena Galić (predsjednik povjerenstva)
Committee memberČaslav Livada (član povjerenstva)
Committee memberKrešimir Romić (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Software Engineering)
(Chair of Visual Computing)
Osijek
Defense date and country2021-10-13, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Program Engineering
Abstract
U ovom radu obrađena je podjela strojnog učenja, osnove dubokog učenja, neuroni i njihova funkcionalnost, što su neuronske mreže, kako se optimiziraju parametri modela, načini sprječavanja prenaučenosti i obrada prirodnog jezika. Kod obrade opisane su tradicionalne metode pred obrade, reprezentacija riječi i arhitekture dubokog učenja namijenjene za obradu. Navedeni su nedostatci povratnih neuronskih mreža i razvijene arhitekture koje su smanjile te nedostatke. U praktičnom dijelu rada riješen je problem razumijevanja teksta tj. klasifikacije teksta. Prilikom rješavanja problema klasifikacije korišten je ULMFiT pristup, koji se temelji na prijenosnom učenju, korištenju diskriminativnih stopa učenja i 1cycle pravila. Istrenirani model postigao je točnost od 91.5%. U svrhu prikaza mogućnosti istreniranog modela napravljeno je jednostavno korisničko sučelje.
Abstract (english)
This paper deals with the division of machine learning, the basics of deep learning, neurons and their functionality, neural networks, how model parameters are optimized, how to prevent overfitting and natural language processing. Within natural language this paper delas with preprocessing, word representation and deep learning architectures for natural language processing. The shortcomings of recurrent neural networks and developed architecture that reduce these shortcomings are presented. Practical part of the paper solves the problem of understanding text i.e. text classification. Problem was solved with ULMFiT approach, based on transfer learning, usage of discriminative learning rates and 1cycle policy. The trained model achieved 91.5% accuracy. In order to show the capability of trained model, a simple user interface was made.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:016691
Study programmeTitle: Professional study programme of Electrical Engineering; specializations in: Power Engineering, Automation, Informatics
Course: Informatics
Study programme type: professional
Study level: undergraduate
Academic / professional title: stručni/a prvostupnik/prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike (stručni/a prvostupnik/prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka elektrotehnike)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2021-11-23 14:06:00