Research on the Application of Deep Learning Technology in Intelligent Dialogue Robots

With the breakthrough of the application of deep learning technology in computer vision and natural language processing tasks, the application of deep learning to intelligent dialogue robots has become a new research hotspot. As a novel way of human-computer interaction, intelligent dialogue robots are becoming one of the entrances of mobile search and services, and are increasingly being paid attention to by enterprises and society. However, there are still some problems in the current dialogue robot, such as insufficient use of multi-modal information and weak emotional expression ability. In this paper, a multi-modal intelligent reply generation model based on seq2seq + attention is proposed by using deep learning algorithm, which can effectively use multi-modal information such as text, picture and video to interact. At the same time, on the basis of considering the contextual content information, the model further integrates the emotional transfer change information of the dialogue text. Experimental evaluation results show that the combination of emotional intelligence can make text response generation more emotionally expressive and more vivid response generation results.