Meta-learning for fake news detection surrounding the Syrian war

ABSTRACT: Summary In this article, we pursue the automatic detection of fake news reporting on the Syrian war using machine learning and meta-learning. The proposed approach is based on a suite of features that include a given article's linguistic style; its level of subjectivity, sensationalism, and sectarianism; the strength of its attribution; and its consistency with other news articles from the same “media camp”. To train our models, we use FA-KES, a fake news dataset about the Syrian war. A suite of basic machine learning models is explored, as well as the model-agnostic meta-learning algorithm (MAML) suitable for few-shot learning, using datasets of a modest size. Feature-importance analysis confirms that the collected features specific to the Syrian war are indeed very important predictors for the output label. The meta-learning model achieves the best performance, improving upon the baseline approaches that are trained exclusively on text features in FA-KES. Highlights • A meta-learning model for distinguishing “true” from “fake” news around the Syrian war• Features capture sectarian discourse and consistency vis-à-vis ground truth• The framework can be extended to detect fake news in other armed conflicts The bigger picture As world and regional powers get more embroiled in the Syrian war, serious questions arise surrounding the credibility of news documenting the facts of war in the decade-long conflict. The spread of fake news around documentation of the war, going beyond plain news bias, not only compromises the integrity of the actual reporting, but also can contribute to psychological warfare that drives the exodus and constant mobility of refugees and hampers humanitarian planning for delivering aid to distraught communities. In this article, we propose meta-learning and machine learning approaches for automatic detection of fake news arising from the Syrian war. Our work reveals the importance of certain features specific to armed conflict in the Middle East, such as sectarian language, and consistency with respect to ground truth from a fact-checking repository. It also highlights the efficacy and quality of meta-learning techniques when tackling datasets of a modest size. We pursue the automatic detection of fake news around the Syrian war using machine learning and meta-learning, using features such as a given article's linguistic style, its level of subjectivity and sectarianism, the strength of its attribution, and its consistency with other articles from the same “media camp”. Our analysis confirms that these features are very important predictors for the output label. Our approach can detect true from fake articles from the FA-KES dataset with extremely high accuracy.

SUBMITTER: Abu Salem F 

PROVIDER: S-EPMC8600244 | biostudies-literature |

REPOSITORIES: biostudies-literature