随着人工智能的不断发展，智能客服、问答等技术越来越受到关注。智能客服系统在一定程度上节省了人力的成本，提高了企业的自动运营效率。然而，对问题缺乏深层的分析与理解，是目前的在线智能客服系统中存在的重要问题。这就导致系统在使用过程中需要很多人工的代价来辅助对问题进行分析，限制了系统的灵活性和适应性，对系统的性能也有一定的影响。本项 目主要针对智能客服中的用户的问题进行分析，包括复杂问句分解、复述、省略句识别及恢复等技术进行研究，以对问题进行更好地分析。智能客服中的答案可以来自于文档、知识库、常问问题集(FAQ)，本文针对问句中的核心实体及核心实体关系识 别、相似度计算、文本蕴含等技术进行研究，以提高答案获取的准确性。

With the development of artificial intelligence, more and more attention has been paid to the technologies of intelligent online-customer service and question answering. The intelligent customer service system saves the cost of manpower to a certain extent and improves the operation efficiency of the enterprise. However, the lack of deep analysis and understanding of the question is an important problem in the current intelligent online-customer service system. This leads to many costs in the process of using the system to assist in the analysis of the problem, which limits the flexibility and adaptability of the system, and also has a certain impact on the performance of the system. This project mainly analyzes user questions in intelligent online-customer service, including complex question decomposition, paraphrasing, ellipsis recognition and recovery and other technologies, so as to make better analysis of the questions. The answer in intelligent customer service can come from document, knowledge base and frequently asked question set (FAQ). In this paper, we study the core entities and core entity relationship recognition, similarity computation, recognition of Textual Entailment and other technologies, in order to improve the accuracy of answer acquisition.