[Paper Review] A Survey of Fake News: Fundamental Theories, Detection Methods, and Opportunities

Xinyi Zhou, Reza Zafarani|arXiv (Cornell University)|Dec 2, 2018

Misinformation and Its Impacts129 references120 citations

TL;DR

This survey assesses fake news from four detection perspectives (knowledge, style, propagation, and sources), links theories from multiple disciplines, and discusses datasets, fact-checking, and intervention opportunities.

ABSTRACT

The explosive growth in fake news and its erosion to democracy, justice, and public trust has increased the demand for fake news detection and intervention. This survey reviews and evaluates methods that can detect fake news from four perspectives: (1) the false knowledge it carries, (2) its writing style, (3) its propagation patterns, and (4) the credibility of its source. The survey also highlights some potential research tasks based on the review. In particular, we identify and detail related fundamental theories across various disciplines to encourage interdisciplinary research on fake news. We hope this survey can facilitate collaborative efforts among experts in computer and information sciences, social sciences, political science, and journalism to research fake news, where such efforts can lead to fake news detection that is not only efficient but more importantly, explainable.

Motivation & Objective

• Define fake news and distinguish related concepts across authenticity, intention, and news status.
• Survey detection methods from four perspectives: knowledge-based, style-based, propagation-based, and source-based.
• Bridge fundamentals theories from social sciences, psychology, and economics to fake news analysis.
• Discuss manual and automatic fact-checking and dataset construction to support detection and intervention.
• Identify open challenges and interdisciplinary research opportunities for explainable fake news detection.

Proposed method

• Organize detection methods into four perspectives: knowledge-based, style-based, propagation-based, and source-based.
• Describe fact-checking processes, including manual expert- and crowd-sourced approaches, and automatic fact-checking via knowledge bases and knowledge graphs.
• Present a unified knowledge representation (SPO triples) and a two-stage automatic fact-checking workflow (fact extraction and fact-checking).
• Summarize and compare existing fact-checking websites and datasets that support ground-truth for fake news analysis.
• Identify relevant fundamental theories and discuss how they can inform explainable, interdisciplinary fake news research.

Experimental results

Research questions

• RQ1 How can fake news be defined in broad vs. narrow terms and how do these definitions impact detection approaches?
• RQ2 What detection strategies exist across knowledge/content, writing style, propagation dynamics, and source credibility?
• RQ3 What role do manual and automatic fact-checking play in scalable fake news detection?
• RQ4 How can interdisciplinary theories inform the design of explainable fake news detection and intervention methods?
• RQ5 What open challenges and research opportunities arise for cross-disciplinary fake news research?

Key findings

• Fake news is best understood through multiple related concepts (deceptive news, false news, disinformation, misinformation, satire, etc.) distinguished by authenticity, intention, and whether it is news.
• There is no universal definition of fake news; broad (false information across information ecosystem) and narrow (intentionally false news published by a news outlet) definitions are presented.
• Manual fact-checking exists as expert-based and crowd-sourced systems, providing ground-truth but facing scalability and bias challenges.
• Automatic fact-checking relies on knowledge representations (SPO triples) and two-stage processes: fact extraction to build knowledge bases/graphs, then fact-checking against these sources.
• A structured survey of detection methods across four perspectives enables analysis of both content and social-medium patterns, including potential for explainable models.
• In addition to classification approaches, the survey emphasizes dataset construction, ground-truth acquisition, and interdisciplinary collaboration as key to advancing fake news research.