Unanswerable Question Correction in Question Answering over Personal Knowledge Base

People often encounter situations where they need to recall past experiences from their daily life. In this paper, we aim construct a question answering system that enables human query over personal knowledge base. Previous works on base focus finding answers for answerable questions. the real world applications, however, people muddle up facts and ask those questions cannot be answered with This work presents novel consisting of model generation model. It not only questions, but also corrects unanswerable if necessary. Our recognizes is inconsistent state suggests can form feasible question. Then, are converted an by For refining question, propose based reinforcement learning (RL) editing mechanism. Experimental results show our proposed effective correcting in answering.