限定领域口语对话系统中的商品属性抽取

按功能或问题域划分,商品属性抽取(product feature mining)在限定领域的对话系统中属于口语语言理解(spoken language understanding, SLU)的范畴。商品属性抽取任务只关注自然文本中描述商品属性的特定部分,它是细粒度观点抽取(fine-grained opinion mining)的一个重要的子任务。现有的商品属性抽取技术主要建立在商品的评论语料上,该文以手机导购对话系统为背景,将商品属性抽取应用到整个对话过程中,增强对话系统应答的针对性。使用基于CBOW (continuous bag of words)语言模型的word2vector(W2V)对词汇的语义层面建模,提出一个针对口语对话的指数型变长静态窗口特征表达框架,捕捉不同距离词语组合的重要特征,使用卷积神经网络(convolutional neural network, CNN)结合词汇的语义和上下文层面对口语对话语料中的商品属性进行抽取。词嵌入模型给出了当前词和所给定的属性类别是否存在相关性的证据,而所提出的特征表达框架则是为了解决一词多义的问题。实验结果表明,该方法取得了优于研究进展中方法的商品属性识别效果。

Product Feature Mining in Restricted Domain Spoken Dialogue System

This paper applies the product feature mining on a dialogue system of a mobile phone recommendation assistant, enhancing the focus of the system during the interaction. CBOW (continuous bag of words) language model is used to represent the sematic clue. A feature framework with exponential elongate static window is introduced to capture the import features among the interactions between words of variant distance. We finally utilize convolutional neural network (CNN) to perform product feature mining task. The word embedding representing sematic clue gives the relation between current word and the product feature, while the feature framework can alleviate the word ambiguity. The experiment shows that our model outperforms the state-of the act methods on product feature mining.