Generating Fact Checking Explanations
This addresses the need for automated, high-quality explanations in fact-checking systems, though it is incremental as it builds on existing veracity prediction methods.
The paper tackles the problem of generating justifications for fact-checking verdicts, which is missing in automated fact-checking systems, and finds that jointly modeling explanation generation with veracity prediction improves system performance and explanation quality.
Most existing work on automated fact checking is concerned with predicting the veracity of claims based on metadata, social network spread, language used in claims, and, more recently, evidence supporting or denying claims. A crucial piece of the puzzle that is still missing is to understand how to automate the most elaborate part of the process -- generating justifications for verdicts on claims. This paper provides the first study of how these explanations can be generated automatically based on available claim context, and how this task can be modelled jointly with veracity prediction. Our results indicate that optimising both objectives at the same time, rather than training them separately, improves the performance of a fact checking system. The results of a manual evaluation further suggest that the informativeness, coverage and overall quality of the generated explanations are also improved in the multi-task model.