This repository contains the implementation of a neural network architecture focused on verifying facts against evidence found in a knowledge base. The architecture can perform relevance evaluation and claim verification. We fine-tuned BERT to codify claims and pieces of evidence separately. An attention layer between the claim and evidence representation computes alignment scores to identify relevant terms between both. Finally, a classification layer receives the vector representation of claims and evidence and performs the relevance and verification classification. Our model allows a more straightforward interpretation of the predictions than other state-of-the-art models. We use the scores computed within the attention layer to show which evidence spans are more relevant to classify a claim as supported or refuted. We use the model to verify facts about COVID-19. The COVID-19 facts corpus is also provided here.
Ramón Casillas (PCIC-UNAM), Helena Gómez-Adorno (IIMAS-UNAM), Victor Lomas-Barrie (IIMAS-UNAM) and Orlando Ramos-Flores (IIMAS-UNAM)