CheckThat! Lab is Back for its Third Version
The 2020’s edition of CheckThat! lab will focus on developing automated systems to identify and verify factual claims in social media. The efforts come as a response to the abundance of information that is being rapidly shared on those platforms. These aspects made it increasingly challenging to distinguish between accurate and misleading content. Thus, an automated system to identify false claims is going to be very welcomed.
QCRI’s Dr. Giovanni Da San Martino and Dr. Preslav Nakov will actively participate in the lab. They will provide an overview of three Arabic tasks that utilized data collected from Twitter.
Task 1 asked the participants to rank the tweets according to their check-worthiness.
Task 2, the participants were asked to rank the verified claims that were identified in Task 1.
Task 3 was evidence retrieval where evidence was collected and used to verify claims.
In Task 4, an expert fact-checker verified the claim’s veracity.
Task 5 is introduced as a legacy task from the previous editions of the lab, and it focused on check-worthiness estimation in debates and speeches.
Dr. Firoj Alam will join Dr. Giovanni Da San Martino and Dr. Preslav Nakov in presenting another paper on the automatic identification of claims in English tweets on Covid-19.
The CheckThat! lab not only aims to help advance the state of the art in natural language processing but also to contribute to the development of tools that can help users navigate the complex landscape of online information more effectively.