MSc Thesis Proposal - "Fake news detection with retrieval augmented generative artificial intelligence" - By: Mohammad Vatani Nezafat

Friday, March 8, 2024 - 11:00 to 12:30

The School of Computer Science is pleased to present…

 

MSc Thesis Proposal by: Mohammad Vatani Nezafat

Title: "Fake news detection with retrieval augmented generative artificial intelligence"

Date: Friday, 08 Mar 2024

Time: 11:00 am – 12:30 pm

Location: Memorial Hall, Room 311

 
Abstract:
This study introduces a novel approach to combating the proliferation of fake news through the integration of MixTral, a state-of-the-art Sparse Mixture of Experts (SMoE) Large Language Model, with a Retrieval-Augmented Generation (RAG) framework utilizing different search APIs for real-time article retrieval. By harnessing MixTral's advanced language understanding capabilities and RAG's dynamic information retrieval process, our system aims to improve the detection and classification of fake news articles significantly. The methodology involves first employing search APIs within the RAG component to fetch contextually relevant information from the web, which is then analyzed alongside the original news content by the MixTral model. This dual-stage process enriches the evaluation context, allowing for a more nuanced and informed assessment of news authenticity. We curated a comprehensive dataset of labeled news articles to train and validate our model, focusing on optimizing the synergy between retrieval and generative mechanisms for superior fake news identification. Preliminary results indicate a marked improvement in detection accuracy over traditional methods, highlighting the potential of combining retrieval-augmented strategies with advanced generative models for misinformation mitigation. This research contributes to the evolving landscape of digital information credibility, offering insights into the development of more effective tools for fake news detection in an era marked by rapid information dissemination and manipulation.
 
Thesis Committee:
Internal Reader: Dr. Luis Rueda 
External Reader: Dr. Mohammad Hassanzadeh  
Advisor: Dr. Saeed Samet
 
MSc AI Seminar - Vector logo