Copenhagen IR lab takes 1st and 2nd place at International Fact Checking competition
The Copenhagen Information Retrieval Lab participated in both tracks of the CLEF fact checking competition and won first prize in the "Factuality" track and second prize in the "Checkworthiness" track.
The CLEF fact checking competition is an international competition for fully automated fact checking software aimed at combating false claims in, e.g., news and politics debates.
In the "Factuality" track, the teams had to check the factuality of a number of identified check-worthy claims. In particular, given a sentence that is worth checking, the goal was for the system to determine whether the claim is likely to be true, half-true, or false, or that it is unsure of its factuality.
In the "Checkworthiness" track, the goal was to predict which claim in a political debate should be prioritised for fact-checking. In particular, given a debate, the goal was to produce a ranked list of its sentences based on their worthiness for fact checking.
The Copenhagen Team
The software that ensured the Copenhagen team the high places in the competition was built by PhD students Casper Hansen, Christian Hansen, and Dongsheng Wang together with their DIKU supervisors Christina Lioma and Jakob Grue Simonsen and collaborator Birger Larsen from Aalborg University. They used solutions based on a combination of methods from information retrieval and artificial intelligence, including convolutional and recurrent neural networks, and support vector machines.
The team competed against several international teams, including participants from the United States, Spain, and India.