面向任务口语对话系统中不含槽信息话语的端到端对话控制

端到端(end-to-end)模型因其能有效避免传统管道式设计存在的错误传递与累积问题,成为了近年来口语对话系统(spoken dialogue system, SDS)的研究热点。在面向任务SDS的end-to-end对话控制中,处理携带任务领域语义信息(槽信息)的话语可以结合命名实体识别、数据库查询结果等语义特征,而不含槽信息的话语,由于缺乏领域语义信息以及表达多样,其有效对话控制仍然是一个挑战。该文提出一种融合“显式”话语特征和“隐式”上下文信息的end-to-end混合编码网络用于处理不含槽信息话语。具体地,在应用卷积神经网络(convolutional neural network, CNN)对“显式”话语序列提取得到的特征表达的基础上,通过构造和捕获对话序列中“隐式”的系统后台上下文信息,进一步丰富了系统动作分类模型的特征表达。在限定领域面向中文任务SDS中的评估结果表明,与传统的管道式SDS和经典的end-to-end SDS相比,该文的方案在不含槽信息话语的单回合处理以及对话段整体性能上都得到了显著提升。

End-to-end Dialogue Control for Utterances without Slot Values in Task-oriented Dialogue System

The end-to-end dialogue control of utterances without slot values is a challenging issue. This paper proposes an end-to-end hybrid coding network that combines explicit utterance features and implicit context information to handle utterances without slot information. Specifically, on the basis of feature expressions extracted from the "explicit" dialogue sequence by the convolutional neural network (CNN), the system action classification model is further enriched by constructing and capturing the "implicit" background system context information in the dialogue sequence. Experiments on the task-oriented restricted domain Chinese SDS shows that, compared to the existing methods, the proposed method achieves significant improvements in both per-response accuracy and per-dialog accuracy.