Given the title of a fake news article A and the title of a coming news article B, program classifies B into agree, disagree, and unrelated.