Research On Reading Comprehension Method Of Automatic Question Answering For Civil Aviation Customer Service

With the continuous expansion of civil aviation,the volume of passenger transportation for civil aviation has increased year by year,and the demand for information from passengers has increased dramatically.The traditional manual customer service of the call center has problems such as high customer service pressure and untimely response,which can no longer meet the development of contemporary civil aviation service industry.How to provide passengers with self-service consulting services to increase user satisfaction has become even more important.The intelligent automatic question answering system can help passengers quickly acquire knowledge from massive civil aviation data,which is the future development trend of call centers.Therefore,it is of great significance to study efficient automatic question answering methods in the field of civil aviation customer service.First,to solve the problem that the traditional question answering system does not have the ability of language understanding,a reading comprehension model based on coattention and adaptive adjustment is proposed.The model uses cosine similarity to adaptively adjust the text vector representation,filter redundant information,learn sentence semantic information with recurrent neural network,introduce coattention mechanism to fuse questions and text information,and use pointer network to extract the answer sequence.Secondly,in response to the problem that users need to filter the required information from massive data in practical applications,this paper combines a fine-grained retriever with a reading comprehension model,and proposes an automatic question answering for civil aviation customer service based on fine-grained retrieval and reading comprehension.The model uses the maximum inner product retrieval algorithm to calculate the relevance of sentence-granularity text to the problem,obtain relevant paragraphs,and then use the reading model to update the problem vector representation and perform iterative retrieval.Finally,the analysis and reasoning of the reading comprehension model are used to generate an answer and return it to the user.To achieve automatic question answering.Finally,using civil aviation data and open domain data for verification,the experimental results show that the model in this paper can quickly and accurately answer passenger questions,and provides a new solution for the automatic question and answer in the field of civil aviation customer service,which is of great significance.