Scholarlink

This chapter introduces the research focus: improving harmful content detection (fake news, hate speech) using transfer learning and data-driven techniques. It highlights the motivation behind studying harmful online content and outlines key challenges in subjective NLP tasks—particularly the scarcity of annotated data and the contextual nature of user-generated content (Sect. 1.1). It presents a set of five guiding research questions addressing preprocessing, contextual modeling, and sentiment integration (Sect. 1.2). This chapter outlines the book’s structure and the experimental goals for each chapter. This chapter concludes by positioning this book within existing literature and summarizing the main contributions, which include novel experiments, enriched data sets, and the integration of affective signals into fact-checking pipelines.

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

Introduction

  • 章节

  • 发表日期：

    2025年

查看全文
  • Salar Mohtaj

摘要

This chapter introduces the research focus: improving harmful content detection (fake news, hate speech) using transfer learning and data-driven techniques. It highlights the motivation behind studying harmful online content and outlines key challenges in subjective NLP tasks—particularly the scarcity of annotated data and the contextual nature of user-generated content (Sect. 1.1). It presents a set of five guiding research questions addressing preprocessing, contextual modeling, and sentiment integration (Sect. 1.2). This chapter outlines the book’s structure and the experimental goals for each chapter. This chapter concludes by positioning this book within existing literature and summarizing the main contributions, which include novel experiments, enriched data sets, and the integration of affective signals into fact-checking pipelines.