Conversational Question Answering with Reformulations over Knowledge Graph

arXiv:2312.17269v2 Announce Type: replace
Abstract: Conversational question answering (convQA) over knowledge graphs (KGs) involves answering multi-turn natural language questions about information contained in a KG. State-of-the-art methods of ConvQA often struggle with inexplicit question-answer pairs. These inputs are easy for human beings to understand given a conversation history, but hard for a machine to interpret, which can degrade ConvQA performance. To address this problem, we propose a reinforcement learning (RL) based model, CornNet, which utilizes question reformulations generated by large …

abstract art arxiv beings conversation conversational cs.ai cs.cl easy graph graphs history human information inputs knowledge knowledge graph knowledge graphs language natural natural language question question answering questions state struggle type