prikaz prve stranice dokumenta Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda
Access restricted to authenticated users
undergraduate thesis
Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda
Split: University of Split, University of Split, Faculty of science, 2020. urn:nbn:hr:166:494514
Galov, Antonio
University of Split
Faculty of Science
Department of Informatics

Institutional repository: Repository of Faculty of Science

Cite this document

Galov, A. (2020). Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda (Undergraduate thesis). Split: University of Split, Faculty of Science. Retrieved from https://urn.nsk.hr/urn:nbn:hr:166:494514

Galov, Antonio. "Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda." Undergraduate thesis, University of Split, Faculty of Science, 2020. https://urn.nsk.hr/urn:nbn:hr:166:494514

Galov, Antonio. "Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda." Undergraduate thesis, University of Split, Faculty of Science, 2020. https://urn.nsk.hr/urn:nbn:hr:166:494514

Galov, A. (2020). 'Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda', Undergraduate thesis, University of Split, Faculty of Science, accessed 04 September 2024, https://urn.nsk.hr/urn:nbn:hr:166:494514

Galov A. Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda [Undergraduate thesis]. Split: University of Split, Faculty of Science; 2020 [cited 2024 September 04] Available at: https://urn.nsk.hr/urn:nbn:hr:166:494514

A. Galov, "Usporedbe algoritama strojnog učenja na problemu dvonožnoga hoda", Undergraduate thesis, University of Split, Faculty of Science, Split, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:166:494514

Metadata
TitleUsporedbe algoritama strojnog učenja na problemu dvonožnoga hoda
Title (english)Comparison of machine learning algorithms for bipedal gait training
AuthorAntonio Galov
MentorGoran Zaharija (mentor)
Committee memberGoran Zaharija (predsjednik povjerenstva)
Committee memberSaša Mladenović (član povjerenstva)
Committee memberDivna Krpan (član povjerenstva)
GranterUniversity of Split
Faculty of Science
(Department of Informatics)
Split
Defense date and country2020-09-29, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Ovaj preddiplomski rad proučava strojno učenje sa fokusom na poduprto učenje i pravi usporedbu tri algoritma (DQN, NEAT i Evostra) na problemu dvonožnoga hoda u Pythonu. Prvo poglavlje priča uvodno o umjetnoj inteligenciji, našem problemu, primjeni u stvarnom životu te cilju završnog rada. Drugo poglavlje priča o strojnom učenju i njegovim vrstama, te detaljno o poduprtom učenju i primjerima njega. Treće poglavlje priča o okolini koju koristimo te o praktičnoj primjeni to jest o svakom od naša tri algoritma objašnjava što su i kako funkcioniraju te kako su implementirani u našem slučaju. Četvrto poglavlje iznosi rezultate testiranja i raspravlja ih. Peto poglavlje daje zaključak usporedbe, razmatra je li poduprto učenje dobro za problem, te govori o tome kako bi algoritmi mogli biti unaprijeđeni.
Abstract (english)
Thesis studies machine learning with focus on reinforcement learning and compares 3 algorithms (DQN, NEAT and Evostra) on the problem of bipedal gait training in Python. First chapter speaks about Artificial Intelligence, given problem, its use in real life, and goal of the thesis. Second chapter goes a bit deeper into machine learning and its types and goes into details about reinforcement learning and examples of it. Third chapter explains te environment we used, and explain what out algorithms are and how they function and lastly how we implemented them. Fourth chapter states results of the testing for all three algorithms and discusses them. Fifth chapter gives a final conclusion of the comparison, discusses if reinforcement learning is the best for given problem, and also argues if the algorithms can be upgraded.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:166:494514
Study programmeTitle: Physics and Computer Science
Study programme type: university
Study level: undergraduate
Academic / professional title: sveučilišni/a prvostupnik/prvostupnica (baccalaureus/baccalaurea) fizike i informatike (sveučilišni/a prvostupnik/prvostupnica (baccalaureus/baccalaurea) fizike i informatike)
Type of resourceText
File originBorn digital
Access conditionsAccess restricted to authenticated users
Terms of useLicence In copyright icon
Created on2021-01-14 10:12:33