Enhanced Natural Language Processing using Transformer Architecture to Detect Fake News and Rumors

Read full text Summarise with VIRAT
Authors: N. Kavitha, S. Lekasri, M. Varsha
Cited by: 0
References: 15

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Nowadays, social networks, online articles, and blogs are the sources for spreading the fake news widely. In social media, all the posts or contents are not original; they may contain fake news or rumors. This fake news can circulate as much as fast and negatively impact the people. This is one of the reasons that we consider that the fake news and rumors detection is more important. So, the fake news detection is necessary to purify the Internet environment. So, we decided to use the transformer architecture model because it is based on neural network architecture and it is designed to handle the stream of data, particularly performing tasks like text generation. The major use of transformer architecture is the self-attention mechanism, which allows the model to focus on different parts of the input sequence when making predictions. The Transformer architecture can be extended to Multi Task Layer (MTL) by modifying the network to handle different tasks. In the MTL model, the transformer architecture learns a shared-attention mechanism across different tasks. Transformer architecture has been widely used in Natural Language Processing (NLP) because the process of NLP can be split into several steps that help to transform raw language data into a form that machines can process and understand.

Fields of Study & Subjects

Misinformation and Its Impacts Spam and Phishing Detection Topic Modeling

Publication details

Pages
129–143
Publication date
1 January 2026
Language
EN
Document type
Book chapter