Detekcija lažnih vijesti pomoću obrade prirodnog jezika

master's thesis

• Title (english): Using Natural Language Processing for Fake News Detection
• Author: Ivana Cvitanović
• Mentor: Marina Bagić Babac (mentor)
• Committee member: Marina Bagić Babac (predsjednik povjerenstva)
• Committee member: Marko Horvat (član povjerenstva)
• Committee member: Krešimir Fertalj (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2021-07-13, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Automatska detekcija lažnih vijesti važan je zadatak u području obrade prirodnog jezika. U ovom radu iznešen je kratki pregled prijašnjih istraživanja i pristupa problemu detekcije lažnih vijesti. Konstruran je skup podataka korištenjem javno dostupnih skupova podataka namjenjenih za detekciju lažnih vijesti. Četri binarna klasifikatora su istrenirana i validirana. Sva četri modela koriste “self-attention” mehanizam. Tri modela koriste tradicionalno nadzirano strojno učenje dok četvrti model za isti zadatak koristi prijenosno učenje. Sva četri modela ostvaruju komparabilne rezultate. Model treniran korištenjem prijenosnog učenje pokazuje najbolju točnost.

Abstract (english)

Automatic fake news detection is an important subject in NLP. This work overviews some of the previous research and approaches on a subject of fake news detection. A dataset is constructed using three publicly available datasets intended for fake news detection. Four different binary classification models are trained and evaluated. All four models utilize the self-attention mechanism. Three models explore traditional supervised learning. The fourth model explores transfer learning, by fine-tuning the pre-trained language model for the same task. All four models yield comparable results with the fourth model achieving the best accuracy.

• Keywords: Obrada prirodnog jezika; lažne vijesti; “self-attention”; duboko učenje; prijenosno učenje
• Keywords (english): Natural Language Processing; Fake News; Self-Attention; Deep Learning; Transfer Learning
• Language: english
• URN:NBN: urn:nbn:hr:168:603927
• 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: 2023-01-13 14:42:08