Flexible, Rapid Authoring of Goal-Orientated, Multi-Turn Dialogues Using the Task Completion Platform

17TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2016), VOLS 1-5: UNDERSTANDING SPEECH PROCESSING IN HUMANS AND MACHINES(2016)

引用 0|浏览91
暂无评分
摘要
The Task Completion Platform (TCP) is a multi-domain, multi modal dialogue system that can host and execute large numbers of goal-orientated dialogue tasks. TCP is comprised of a task configuration language, TaskForm, and a task-independent dialogue runtime, allowing task definitions to be decoupled from the global dialogue policy used by the platform to execute the tasks. This separation enables scenario developers to rapidly develop new dialogue systems, by eliminating the need to re implement the policy from scratch for each new task. In this paper, we introduce support for authoring tasks in a variety of dialogue styles, ranging from entirely flexible to fully system initiative. This flexibility is enabled by a set of task-level policy override constructs, which augment or constrain the default platform-level policy to achieve the desired system behavior. We demonstrate the use of the TaskForm language to define complex, multi-turn tasks in a variety of domains and add different task-specific policy constructs to demonstrate the flexibility of the task authoring process.
更多
查看译文
关键词
dialog systems, dialog policy
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要