Statistical methods for building robust spoken dialogue systems in an automobile
Statistical methods for building robust spoken dialogue systems in an automobile
复制标题
在汽车中构建强大的语音对话系统的统计方法
DOI：
--
复制
发表时间：
2012
期刊：
影响因子：
--
通讯作者：
E. Tzirkel
中科院分区：
文献类型：
--
作者：
Pirros Tsiakoulis;M. Gašić;Matthew Henderson;J. Planells;J. Prombonas;Blaise Thomson;Kai Yu;S. Young;E. Tzirkel
©2013 by Taylor & Francis Group， LLC.我们研究了统计技术在汽车环境中口语对话系统的潜力。具体来说，我们关注部分可观察的马尔可夫决策过程（pomdp），它最近被提议作为构建对话管理器（dm）的统计框架。这些统计dm具有明确的不确定性模型，允许利用替代识别假设，以及可以使用强化学习自动优化的对话管理策略。本文介绍了一种基于语音的车载系统，用于提供有关当地便利设施（如餐馆）的信息。本文描述了一个用户试验，该试验比较了经过训练的统计对话管理器与传统手工制作系统的性能。结果显示了两种系统的不同行为以及使用统计方法获得的性能优势。
© 2013 by Taylor & Francis Group, LLC. We investigate the potential of statistical techniques for spoken dialogue systems in an automotive environment. Specifically, we focus on partially observable Markov decision processes (POMDPs), which have recently been proposed as a statistical framework for building dialogue managers (DMs). These statistical DMs have explicit models of uncertainty, which allow alternative recognition hypotheses to be exploited, and dialogue management policies that can be optimised automatically using reinforcement learning. This paper presents a voice-based in-car system for providing information about local amenities (e.g. restaurants). A user trial is described which compares performance of a trained statistical dialogue manager with a conventional handcrafted system. The results demonstrate the differing behaviours of the two systems and the performance advantage obtained when using the statistical approach.