Our main goal is to develop fact checking algorithms that can assess the credibility of claims mentioned in the textual statements and provide interpretable valid evidence that explains why a certain claim is considered as factually true or fake.
Fact-checking in textual statements aims to identify claims expressed in statements and label them as factually true or false. Fact-checking has significant real-world impacts as it enables debunking false claims and supporting the true ones. However, it is a challenging and time-consuming task for humans as, for example, journalists/editors have to fact-check information and identify reliable evidences from various sources manually.
In this project, we attempt to automate the process of reasoning for fact-checking. In particular, we aim to (a) automatically provide an interpretable valid set of evidences for a given textual-statement/claim from reliable data sources, (b) develop effective and efficient fact checking algorithms that can reason about their predictions by assessing the veracity of textual statements using provided evidences, and (c) develop resources and datasets to evaluate our algorithms in different genres including news about celebrities and politicians, as well as crowdsourced deceptive statements.