Towards Improving the Performance of Question Answering System using Knowledge Graph - A Survey

A question answering system attempts to give a precise response to questions posed in natural language. Question answering systems must capture some context information in order to accurately answer non-trivial queries. Question answering is a challenging research issue because of the contextual interconnected information. Knowledge graphs are a form of data representation that is both structured and interconnected. The process of retrieval of exact answer from a knowledge graph for a given question is named as knowledge graph question answering (KGQA). Different ways for executing KGQA exist depending on the type of query, searching patterns over the knowledge graph, ranking of candidate answers, and so on. This paper surveys various approaches for KGQA, the major challenges, the current techniques for evaluations and the research gaps in KGQA.