Primjena podržanog učenja kod igara

undergraduate thesis

• Title (english): Reinforcement Learning Applied to Games
• Author: Mateo Puhalović
• Mentor: Tomislav Pribanić (mentor)
• Committee member: Tomislav Pribanić (predsjednik povjerenstva)
• Committee member: Kristijan Bartol (član povjerenstva)
• Committee member: David Bojanić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2019-09-04, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Podržano učenje je jedna kategorija strojnog učenja, uz nadzirano i nenadzirano učenje, te je ono najprikladnije za primjenu kod igara. U podržanom učenju, agent promatra svoju okolinu te uči donositi odluke o akcijama koje treba napraviti kako bi maksimizirao nagradu. Proces na kojem je bazirano podržano učenje naziva se Markovljev proces odlučivanja. U ovom radu korištena je metoda dubokog Q učenja koja kombinira duboko učenje te Q učenje tako da neuronska mreža zamjenjuje Q tablicu. U ovom radu principi podržanog učenja metodom dubokog Q učenja prikazani su na primjeru agenta koji uči igrati igricu „Snake“. Za izradu neuronske mreže korištena je biblioteka Keras.

Abstract (english)

Reinforcement learning is a category of machine learning, along with supervised and unsupervised learning and it is most suitable for application in games. In reinforcement learning, agent observeres it's environment and learns to make decisions about taking actions to maximize it's reward. Process on which reinforcement learning is based is called Markov Decision Process. Method used in this paper is deep Q learning which combines deep learning and Q learning by replacing Q table with neural network. In this paper, principles of reinforcement learning with deep Q learning method is demonstrated on agent that learns to play game Snake. Library that is used for making neural networks is Keras.

• Keywords: Umjetna inteligencija; strojno učenje; podržano učenje; Q učenje; duboko učenje; Markovljev proces odlučivanja; neuronske mreže; Keras
• Keywords (english): Artificial inteligence; machine learning; reinforcement learning; Q learning; deep learning; Markov Decision Process; neural networks; Keras
• Language: croatian
• URN:NBN: urn:nbn:hr:168:769026
• 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: 2020-01-07 21:38:58