Memory network based Knowledge Driven Model for Response Generation in Dialog System

Abstract

Human-machine conversation is one of the most important topics in artificial intelligence (AI) and has received much attention across academia and industry in recent years. Currently dialogue system is still in its infancy, which usually converses passively and utters their words more as a matter of response rather than on their own initiatives, which is different from human-human conversation. This paper tackles the problem of generating informative responses by integrating knowledge base into the dialogue system’s response generation process, in an end-to-end way. A novel architecture is proposed, namely Memory network based Knowledge Driven Model (MKDM), which can integrate knowledge base by memory manager, and generate knowledge grounded responses. By conducting comparative experiments on automatic metrics demonstrate the effectiveness and usefulness of our model.