一种改进的主动对话方法实现

An Improved Implementation of Proactive Human-Machine Conversation

主动对话是对话系统领域内新提出的一个任务,它与传统对话系统的不同在于它可以通过给定的目标来进行有目的的主动对话。通过分析主动对话任务的特点,从编码和融合两个方面对模型进行提升。实验结果表明,在相同的数据集上,所提出的改进的模型比之前的方法具有更好的效果。

Proactive Human-Machine Conversation is a new task in the field of dialogue system.It is different from the traditional dialogue system in that it can carry out proactive dialogue through a given goal.By analyzing the characteristics of this dialogue task,this paper improves the model from two aspects of encoding and fusion.The experimental results show that the improved model proposed has better performance than the previous method on the same data set.