Where is the supporting evidence?

In the field of fact verification, gathering supporting evidence is crucial and poses several challenges. James Thorne, in a discussion on the NLP Highlights, explains that verifying the truthfulness of claims often involves identifying credible sources, whether they are textual claims, triples, or entire documents. This process demands not only retrieving relevant information but ensuring the diversity and trustworthiness of the evidence to validate or refute a claim effectively 1.

For example, in the context of the Fake News Challenge, the data provided in pairs aids systems in classification exercises. However, Thorne emphasizes that the real challenge lies in reasoning about what information is needed to verify a source, covering both correct and incorrect information. This is akin to identifying news article citations for statements in Wikipedia, a process that underscores the importance of finding the right citations to support claims reliably 1.