This paper investigates the impact of word classing on a recently proposed shrinkage-based language model, Model M [5]. Model M, a class-based n-gram model, has been shown to significantly outperform word-based n-gram models on a variety of domains. In past work, word classes for Model M were induced automatically from unlabeled text using the algorithm of [2]. We take a closer look at the classing and attempt to find out whether improved classing would also translate to improved performance. In particular, we explore the use of manually-assigned classes, part-of-speech (POS) tags, and dialog state information, considering both hard classing and soft classing. In experiments with a conversational dialog system (human-machine dialog) and a speech-to-speech translation system (human-human dialog), we find that better classing can improve Model M performance by up to 3% absolute in word-error rate.
展开▼
机译：本文研究了单词分类对最近提出的基于收缩的语言模型Model M [5]的影响。模型M是基于类的n-gram模型，已显示出在许多领域上明显优于基于单词的n-gram模型。在过去的工作中，使用[2]的算法从未标记的文本中自动归纳出模型M的单词类别。我们仔细研究分类，并尝试找出改进的分类是否还会转化为改进的性能。特别是，我们在考虑硬分类和软分类的情况下，探索了如何使用手动分配的类，词性（POS）标签和对话状态信息。在对话式对话系统（人机对话）和语音对语音翻译系统（人-人对话）的实验中，我们发现更好的分类可以将Model M的性能提高高达3％的绝对绝对错误率。
展开▼
