The School of Computer Science is pleased to present...

Knowledge Informed Fake News Detection using Large Language Models

 

MSc Thesis Defense by: Jess Joseph Benny

 

Date: Thursday December 7th, 2023

Time:  2:00 PM – 4:00 PM

Location: Dillon Hall, Room 264

 

Abstract:

The spread of false or misleading information as news has been a significant threat to governments, organizations and the economy for a long time. However, it has become more prevalent and influential in recent years due to the growing popularity of social media, which is now the primary source of information for more than half of the world’s population. Detecting fake news used to rely mostly on statistical and linguistic analysis of texts, but with the advancement of AI and computer-assisted writing, fake news authors can now deceive statistical models. Therefore, now we have more sophisticated methods that can capture the context and knowledge behind fake news, known as knowledge-informed news classification methods.

This thesis proposes a new knowledge-informed fake news detection method that combines knowledge graphs (KG) with large language models (LLMs). Our proposed methodology utilizes KGs to generate knowledge graph embeddings enriched with external knowledge relevant to the subject matter of the news article. Concurrently, LLMs are utilized to generate context-aware, knowledge-enriched document embeddings from the body of the news article. We tried to evaluate the efficacy of our proposed methodology by experimenting with 21 unique combinations of different LLMs (T5, GPT and Ernie) and KG embedding techniques (TransE, RotatE, SimplE).  Our experiments revealed that combining LLMs with KGs demonstrated impressive performance on our evaluation datasets, especially the combination of the T5 language model and SimplE KG embedding technique outperforming the state-of-the-art model on the FakeNewsNet dataset.

Keywords: Fake News, Large Language Models, Knowledge Graphs, NLP, Deep Learning

 

Thesis Committee:

Internal Reader: Dr. Ikjot Saini   

External Reader: Dr. Yuntong Wang        

Advisor: Dr. Dan Wu

Chair: Dr. Muhammad Asaduzzaman