arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 1803.03232cs.CLcs.AIcs.NE
收藏

面向大领域对话管理的封建强化学习

Feudal Reinforcement Learning for Dialogue Management in Large Domains

University of Cambridge(剑桥大学);PolyAI Limited(PolyAI有限公司)

Iñigo Casanueva, Paweł Budzianowski, Pei-Hao Su, Stefan Ultes, Lina Rojas-Barahona, Bo-Hsiang Tseng, Milica Gašić

摘要:

强化学习(RL)是解决对话策略优化的一种很有前途的方法。然而,传统的强化学习算法由于维数灾难而无法扩展到大的域。我们提出了一种新的对话管理架构,基于封建RL,它分解成两个步骤的决定;第一步,主策略选择一个子集的原始动作,和第二步,其中一个原始动作是从选定的子集中选择。包含在领域本体中的结构信息用于抽象对话状态空间,在每个步骤中使用抽象状态的不同部分进行决策。这与插槽之间的信息共享机制相结合,增加了大型域的可扩展性。我们表明,基于Deep-Q网络的这种方法的实现在几个对话领域和环境中显着优于以前的最先进技术,而不需要任何额外的奖励信号。

英文摘要:

Reinforcement learning (RL) is a promising approach to solve dialogue policy optimisation. Traditional RL algorithms, however, fail to scale to large domains due to the curse of dimensionality. We propose a novel Dialogue Management architecture, based on Feudal RL, which decomposes the decision into two steps; a first step where a master policy selects a subset of primitive actions, and a second step where a primitive action is chosen from the selected subset. The structural information included in the domain ontology is used to abstract the dialogue state space, taking the decisions at each step using different parts of the abstracted state. This, combined with an information sharing mechanism between slots, increases the scalability to large domains. We show that an implementation of this approach, based on Deep-Q Networks, significantly outperforms previous state of the art in several dialogue domains and environments, without the need of any additional reward signal.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑