Duboko učenje nad dinamičkim grafovima

master's thesis

• Title (english): Deep learning on dynamic graphs
• Author: Daniel Vusić
• Mentor: Boris Vrdoljak (mentor)
• Committee member: Boris Vrdoljak (predsjednik povjerenstva)
• Committee member: Mirta Baranović (član povjerenstva)
• Committee member: Mario Brčić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Duboko učenje nad dinamičkim grafovima uključuje razne načine rješavanja zadataka koji su postavljeni nad skupom podataka predstavljenim grafom u kontinuiranom vremenu. Klasifikacija čvorova, predikcija veza te još neki zadatci sasvim su novi i specifični izazovi koje metode klasičnog dubokog učenja ne mogu riješiti zbog različitosti podataka. Graf osim značajki čvorova sadrži puno više informacija na temelju njegove topologije, te još više informacija kada se uzme u obzir kako se taj graf oblikovao kroz vrijeme. Modeli dubokog učenja nad grafovima raznim tehnikama koriste upravo te informacije za dobivanje najboljih rezultata.

Abstract (english)

Deep learning over dynamic graphs involves various ways of solving tasks that are set over a set of data represented by a graph in continuous time. Node classification, link prediction and few other tasks are completely new and specific challenges that the methods of classical deep learning cannot solve due to the diversity of data. In addition to the node features, the graph contains much more information based on its topology, and even more information when you consider how that graph was formed over time. Deep learning models of graphs using various techniques use this information to get the best results.

• Keywords: duboko učenje; dinamički grafovi; neuronska mreža; vremenski grafovi; predikcija; koder-dekoder; vektorski prikaz
• Keywords (english): deep learning; dynamic graphs; neural network; time graphs; prediction; encoder-decoder; vector embedding
• Language: croatian
• URN:NBN: urn:nbn:hr:168:926296
• 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 10:16:51