Fact Checking using Knowledge Bases
Description
Full text not available
Abstract
In today’s society, the spread of incorrect information is becoming an increasingly serious issue. Verifying information using unstructured data, such as a text corpus, can be difficult for all claims since they lack semantic understanding and are difficult to justify. As a result, it is important to validate the claim using a knowledge base that can offer logic and is trustworthy. In this thesis, we suggest a baseline and advanced method to verify claims from a knowledge base to prevent the problem of lack of semantic knowledge. Pre-trained deep learning models, like T5 are fine-tuned to verify the claims using knowledge bases. Lastly, analysis and results obtained from the predictions madeby the models are discussed.