Knowledge Graphs, Large Language Models, and Hallucinations: An NLP Perspective