Generiranje klasične glazbe pomoću Restricted Boltzmann machines

undergraduate thesis

• Title (english): Generating classical music with Restricted Boltzmann machines
• Author: Ivan Landeka
• Mentor: Siniša Srbljić (mentor)
• Committee member: Siniša Srbljić (predsjednik povjerenstva)
• Committee member: Adrian Satja Kurdija (član povjerenstva)
• Committee member: Miroslav Popović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-09-03, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U ovom radu obrađuje se nenadzirano učenje u vidu generiranja glazbenih nizova. Na početku rada opisujemo kako je nastao interes za konektivističkim pristupom umjetnoj inteligenciji. Navode se razni modeli neuronskih mreža i gdje se koriste te kako se provodi učenje neuronskih mreža. Posebno se analizira generativni model RBM i metode Monte Carlo kojima generiramo rekonstrukciju podataka. Detaljno je objašnjen skup podataka, implementacija mreže i program za oblikovanje primjera za učenje. Proveli smo treniranje i analizirali rezultate za razne parametre mreže.

Abstract (english)

In this bachelor thesis we study unsupervised learning on generating classical music pieces. In the beginning we describe the rise of interest for connectivistic approach to artificial intelligence. Various models of neural networks are described, when we use them and how to train them. Especially, we analyze generative model RBM and Markov Chain Monte Carlo methods we use for reconstruction of the data. The dataset, implementation of the neural net and modelling the training examples are described in detail. We trained the model on various parameters and explained what the model has learned.

• Keywords: neuronske mreže; umjetna inteligencija; rbm; uzorkovanje; klasična glazba
• Keywords (english): neural networks; artificial intelligence; rbm; sampling; classical music
• Language: croatian
• URN:NBN: urn:nbn:hr:168:507641
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 22:13:55