Description
An international competition, supported by the Association for Computational Linguistics, on the task of automated fact-checking.
My entry, reported in the image as "TheEarthIsFlat", consisted in the implementation of Universal Sentence Encoder to a set of different datasets in order to obtain a ranking of sentences, extracted from political debates, based on their probability of being worthy of fact-checking.
The results are very positive (obtained a second place and the best score in a secondary run) and were published in a paper that described my contributions in the challenge, which has been published by CEUR-WS.
Link Below