Scholarlink

<p>The proliferation of manipulated content, such as counterfeit films, text, audio, and photographs, has surged in recent years due to advanced digital manipulation tools and techniques. Social media platforms are also plagued by false information-laden tweets that can rapidly spread and influence public opinion. Recent advancements in natural language processing have empowered deep neural models with unprecedented generative capabilities, enabling the generation of realistic text content. Unfortunately, adversaries have capitalized on these technological improvements to deploy social bots that disseminate deepfake content, which skews public discussions. Consequently, detecting machine-generated content on sites like Twitter is critical to mitigating this challenge. This work presents a novel approach that integrates deep learning algorithms with word embeddings to distinguish tweets as human-generated or bot-generated. The study introduces a network of Attention Siamese Network (ASN) that successfully classifies the sentiment of deepfake tweets with an accuracy of 0.9825, outperforming existing approaches. The proposed model’s performance was validated using a publicly available Tweepfake dataset. The goal of this research is to improve the automated system’s detection capabilities and curb the spread of deepfake content on social media platforms.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel approach to identify deepfake text using social media data

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月25日

查看全文
  • Battula Thirumaleshwari Devi,
  • Rajkumar Rajasekaran

摘要

The proliferation of manipulated content, such as counterfeit films, text, audio, and photographs, has surged in recent years due to advanced digital manipulation tools and techniques. Social media platforms are also plagued by false information-laden tweets that can rapidly spread and influence public opinion. Recent advancements in natural language processing have empowered deep neural models with unprecedented generative capabilities, enabling the generation of realistic text content. Unfortunately, adversaries have capitalized on these technological improvements to deploy social bots that disseminate deepfake content, which skews public discussions. Consequently, detecting machine-generated content on sites like Twitter is critical to mitigating this challenge. This work presents a novel approach that integrates deep learning algorithms with word embeddings to distinguish tweets as human-generated or bot-generated. The study introduces a network of Attention Siamese Network (ASN) that successfully classifies the sentiment of deepfake tweets with an accuracy of 0.9825, outperforming existing approaches. The proposed model’s performance was validated using a publicly available Tweepfake dataset. The goal of this research is to improve the automated system’s detection capabilities and curb the spread of deepfake content on social media platforms.