Comprehensible Context-driven Text Game Playing
Comprehensible Context-driven Text Game Playing
复制标题
易于理解的上下文驱动的文本游戏
DOI：
--
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Jonathan May
中科院分区：
文献类型：
--
作者：
Xusen Yin;Jonathan May
为了训练一个计算机智能体玩一个基于文本的计算机游戏，我们必须表示游戏的每个隐藏状态。在观察到的文本上运行的长短期记忆（LSTM）模型是状态构建的常见选择。然而，这种智能体的正常深度Q学习网络（DQN）需要数百万步的训练或更多步骤才能收敛。因此，基于LSTM的DQN可能需要数十天才能完成训练过程。虽然我们可以使用卷积神经网络（CNN）作为文本编码器来比LSTM更快地构建状态，但在不理解所分析单词的语法上下文的情况下这样做可能会减慢收敛速度。在本文中，我们使用快速CNN来编码从观察到的文本中提取的面向位置和语法的结构作为状态。我们还以通用和实用的方式增强了奖励信号。总之，我们表明，我们的改进不仅可以加快一个数量级的过程，但也学习一个上级代理。
In order to train a computer agent to play a text-based computer game, we must represent each hidden state of the game. A Long Short-Term Memory (LSTM) model running over observed texts is a common choice for state construction. However, a normal Deep Q-learning Network (DQN) for such an agent requires millions of steps of training or more to converge. As such, an LSTM-based DQN can take tens of days to finish the training process. Though we can use a Convolutional Neural Network (CNN) as a text-encoder to construct states much faster than the LSTM, doing so without an understanding of the syntactic context of the words being analyzed can slow convergence. In this paper, we use a fast CNN to encode position-and syntax-oriented structures extracted from observed texts as states. We additionally augment the reward signal in a universal and practical manner. Together, we show that our improvements can not only speed up the process by one order of magnitude but also learn a superior agent.