MEX-A3T: Fake News and Aggressiveness Analysis

E-communication in general, and social networks, in particular, are increasingly playing crucial roles in everyone's life. Because of that, the analysis of textual information coming from social networks has been a popular research topic among the computational linguistics community. In this sense, very effective methods have been developed for such purpose, resulting in a better understanding on how to deal with inherent problems from such domain, such as shortness, slang, non-thematic nature, multilingualism, multimodality, among others. Largely, this research progress can be attributed to academic competitions or dedicated tasks that seek to advance the state of the art in a particular research topic of practical relevance (see e.g., the series of events organized by IberEval , TASS and PAN ).

Despite of such progress, there are still open issues that deserve further research in order to be solved or at least to better understand them. Accordingly, in the previous year we organized a shared task at IberLEF 2019 aimed to advance the state of the art on the non-thematic analysis of short texts written in Mexican Spanish. In particular, the 2019 edition of MEX-A3T considered two main tracks: on the one hand, a track on multi-modal author profiling, whose aim was to develop methods for profiling users according to non-standard dimensions (gender, occupation and place of residence), and, on the other hand, a track on aggressiveness detection in tweets.

The goal of the third edition of MEX-A3T is to further improve the research in NLP tasks as well as to continue pushing the computational treatment of the Mexican Spanish. As a novelty, this year’s proposal introduces a new track on fake news detection and an improved corpus for the offensive language detection track. The MEX-A3T@IberLEF2020 has the following two tracks:

FAKE NEWS TRACK: Fake news provides information that aims to manipulate people for different purposes: terrorism, political elections, advertisement, satire, among others. In social networks, misinformation extends in seconds among thousands of people, so it is necessary to develop tools that help control the amount of false information on the web. Similar tasks are detection of popularity in social networks and detection of subjectivity of messages in this media. A fake news detection system aims to help users detect and filter out potentially deceptive news. The prediction of intentionally misleading news is usually based on the analysis of truthful and fraudulent previously reviewed news, i.e., annotated corpora. The Fake News track consists in classifying a given set of news written in Mexican Spanish between true and fake.

AGGRESSIVENESS DETECTION TRACK: This track follows up on last year's evaluation task; it focuses on the detection of aggressive tweets in Mexican Spanish. It is important to mention that for this year the criteria for identifying aggression have been revised and a new data set has been created. The new dataset is similar in size and class ratio to previous years.