LLEAD – Layered Learning for Early Anomaly Detection: A Metalearning Approach

Description:

While social media have the potential to spread misinformation, they also have the potential to support the education of the public. The goal of this project is to develop a bot that participates in debates in social media by detecting misinformation using contradiction detection methods and providing reliable information, and to test it in discussions about COVID-19.

 

Outcome:

- adaptation of state-of-the-art approaches for contradiction detection to the problem of detecting misinformation;
- empirical comparison of the methods developed on COVID-19 data;
- a bot that uses the methods developed to participate in social media debates;
- publication of a paper.

 

Author: Tomás Novo

Type: MSc thesis

Partner: FEUP – Faculdade de Engenharia da Universidade do Porto

Year: 2021