知识驱动的人机主动对话管理策略

针对当前的对话系统主要是被动响应回复，尚不能较好地进行主动式对话的问题，提出了一种知识驱动的人机主动对话管理策略，模拟人-人交流模式，将对话分为话题切换和话题深入两个子任务，设计个性化对话管理策略来实现多轮对话中的主动引导和话题转移。该策略依据人机交互情感状态确定系统主动对话时机，利用知识图谱作为背景知识信息，主动搜索其在知识图谱中触发的对话实体的多跳邻居集合，从而决定下一步交互内容。针对用户情感消极的话题，通过向外传播法来主动寻找新话题；针对用户情感积极的话题，通过向内聚合法来深入响应当前话题。实验结果表明，该策略在平衡全局对话连贯性和局部话题一致性的同时，提高了系统对话的主动性，为人机主动对话系统的发展提供了新的参考。

Knowledge Driven Management Strategy of Human-Machine Active Dialogue

To solve the problem that the current dialogue system is mainly passive response and stillunable to carry out active dialogue well, a knowledge driven human-machine active dialogue managementstrategy is proposed, which simulates human communication mode and divedes the dialogue into two sub-tasks: topic switching and topic depth. A personalized dialogue management strategy is designed torealize active guidance and topic transfer in multi-round dialogues. The proposed strategy determines thetime of the system's active dialogue based on the emotional state of human-machine interaction, and usesthe knowledge graph as the background knowledge information to actively search the multi-hop neighborset of the dialogue entities that are triggered by the knowledge graph, so as to determine the nextinteraction content. For topics of users' negative emotions, new topics are actively sought for throughoutward communication method. For topics with users' positive emotions, the current topic can be deeplyrespond to through cohesion. The experimental results show that the initiative of model dialogue isimproved by the strategy while balancing global dialogue coherence and local topic consistency, which isa new reference for the development of human-machine active dialogue system.