Paper Contents
Abstract
The spread of fake news in recent years has presented serious obstacles to public trust and information integrity. Sophisticated methods that can tell the difference between fake and real content are needed to solve this problem. In order to identify bogus news, this review paper investigates the use of transfer learning in combination with multi-model neural networks. Transfer learning improves the effectiveness and precision of fake news detection systems by using pre-trained models to adapt knowledge from one domain to another. The first section of the paper provides an outline of false news, emphasising its effects on society and the challenges associated with identifying it. After that, it explores the fundamentals of transfer learning, highlighting how learnt characteristics from big datasets can be transferred to optimise neural network performance. We explore different approaches and architectures used in multi-model neural networks to show how adding textual, visual, and social context together with other modalities improves the resilience of fake news detection systems. The article also examines case studies and recent developments that have effectively used transfer learning to enhance the detection of false news across various platforms and languages. Important issues in this quickly developing subject are covered, including adversarial assaults and dataset biases, as well as new trends and possibilities for future research.As a conclusion, this review offers a roadmap for future developments in this important field of research by synthesising existing knowledge and shedding light on the possible applications of multi-model neural networks and transfer learning to stop the spread of false information.
Copyright
Copyright © 2024 AMRIT SINGH. This is an open access article distributed under the Creative Commons Attribution License.
