Federalno učenje nad tokovima podataka

master's thesis

• Title (english): Federated Learning Over Data Streams
• Author: Luka Namačinski
• Mentor: Igor Čavrak (mentor)
• Committee member: Igor Čavrak (član povjerenstva)
• Committee member: Ivana Bosnić (član povjerenstva)
• Committee member: Danko Basch (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-01, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Rad daje pregled područja federalnog učenja. Predstavlja motivaciju i okolnosti koje su dovele do njegovog razvoja. Nadalje opisujemo ključne izazove s kojima se federalno učenje suočava. Također prikazuje korisnu taksonomiju za kategorizaciju i analizu sustava za federalno učenje. Predlaže sustav za federalno učenje implementiran u programskom jeziku Python i sustavu za razmjenjivanje poruka Apache Kafka. Objašnjava motivaciju odabira konkretnih tehnologija i način na koje su integrirane u sustav. Konačno, prikazujemo kratku analizu performansi izgrađenog sustava. Sustav je testiran za preciznost i skalabilnost sa zaključkom da postiže zadovoljavajuće performanse u obje kategorije.

Abstract (english)

This thesis presents an overview of federated learning. It shows the circumstances and motivation which led to its development as well as the challenges that federated learning faces. We present a categorization of federated learning, useful for analyzing a given federated learning system. Furthermore, we present an implemented federated learning system implemented in Python using the distributed messaging system Apache Kafka. Moreover, we present the reasoning behind the selection of these technologies and how they were incorporated into the system. Finally, we analyze the performance of the created system. Model quality and scalability were tested with the results showing that the system performed well

• Keywords: federalno učenje; tokovi podataka; strojno učenje; privatnost; Apache Kafka
• Keywords (english): federated learning; data streams; machine learning; privacy; Apache Kafka
• Language: croatian
• URN:NBN: urn:nbn:hr:168:242309
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-01-14 09:35:07