Prova de Doutoramento da aluna Danielle Caled Vieira
Área: Engenharia Informática e de Computadores
Título da Tese: A data science approach for handling misinformation in digital media
Local da Prova: https://videoconf-colibri.zoom.us/j/92621116358
Data: 16/01/2024
Hora: 16h30
Abstract: Massive online textual content emerges every day through multiple channels, also bringing with it misinformation. Fact-checking initiatives on news and social media are unable with the current scale and diversity of misinformation. Rather than attempting to label/filter misleading information, this thesis explores how to empower news consumers with assisting tools designed to facilitate the assessment of the credibility of textual content. This is achieved through the development of an article assessment resource bundle, composed by a multidimensional indicator, a set of explanatory metrics, a package for validating the article's source, and an explainer, which generates a short text summarizing the results of the indicator and metrics. The multidimensional indicator is designed as the classification result of an input online article as hard news, opinion, soft news, satire, or conspiracy theory. Unlike a binary classification, this indicator provides a more comprehensive evaluation of an article. Explanatory metrics supporting the classification results include are the subjectivity, sentiment, headline representativeness, headline sensationalism, and lexical coverage. The article assessment resource bundle is analyzed from the perspective of a news consumer. In particular, the analysis is centered on the evaluation of the multidimensional indicator and explanatory metrics to understand: i) to what extent a computer-generated assessment may affect news consumers' perceptions of the credibility of an article credibility, and how this influence is manifested; ii) the effectiveness of automatic article classification in aiding news consumers to assess credibility; and iii) the most relevant explanatory metrics to support the conscious consumption of news content.
