Title: Countering Misinformation in Data-driven Narratives through Automated Fact-checking and Data Visualization

Date: Monday, May 13, 2024

Time: 9:00 AM - 11:00 AM EST

Location: TSRB 334 (VIS Lab)

Virtual: Zoom link

Yu Fu

Human-centered Computing Ph.D. Student

School of Interactive Computing

Georgia Institute of Technology

Committee

Dr. John Stasko (advisor) – School of Interactive Computing, Georgia Institute of Technology

Dr. Alex Endert – School of Interactive Computing, Georgia Institute of Technology

Dr. Cindy Xiong Bearfield – School of Interactive Computing, Georgia Institute of Technology

Dr. Munmun De Choudhury – School of Interactive Computing, Georgia Institute of Technology

Dr. Nicholas Diakopoulos – School of Communication, Northwestern University

Abstract

Misinformation threatens media credibility and influences public opinion — the mix of genuine information with falsehoods challenges people to tease out trustworthy content from a web where facts and fallacies are interwoven. Data-driven narratives (i.e., claims or interpretations drawn from structured data and statistics) are often perceived as more trustworthy. Yet, they are vulnerable to inaccuracies and flawed reasoning, which can transform them into misinformation.

Automated Fact-Checking (AFC) holds the potential to enhance fact-checking efforts and more broadly assist readers in evaluating the information they encounter. However, current AFC work on data-driven narratives remains inadequate, with considerable limitations and challenges across various aspects. Furthermore, there has been insufficient research on the effective communication of corroborating justifications or data evidence, which is crucial for making AFC results more explainable and actionable.

My proposed thesis focuses on data-driven misinformation, seeking to enhance our understanding of this issue and develop advanced technologies to address its verification and communication. To achieve this goal, my work first examines the problems undermining the integrity of data-driven narratives, establishing a comprehensive taxonomy and outlining implications for their verification and communication. Next, I identify the downstream tasks associated with its fact-checking and formulate a framework to organize these tasks effectively. I plan to utilize Large Language Models (LLMs) and develop computational models for performing verification tasks, as well as design methods and techniques to communicate the verification results and underlying data evidence. Ultimately, I will integrate the taxonomy, techniques, and designs into interactive systems that cater to the needs of various stakeholders and assess their impact. This work aims to contribute a comprehensive taxonomy, design methodologies, computational models, and interactive prototypes to the field of automated fact-checking, information visualization, and human-computer interaction (HCI).