arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2010.12919cs.CLcs.AI
收藏

语言属性的因果效应

Causal Effects of Linguistic Properties

University of Chicago(芝加哥大学);Columbia University(哥伦比亚大学)

Reid Pryzant, Dallas Card, Dan Jurafsky, Victor Veitch, Dhanya Sridhar

摘要:

我们考虑使用观测数据来估计语言属性的因果效应的问题。例如,礼貌地写一封投诉信是否会导致更快的响应时间?一个正面的产品评价会增加多少销售额?本文解决了两个技术挑战,开发一个实用的方法之前的问题。首先,我们将兴趣的因果量形式化为作者意图的影响,并建立必要的假设来从观察数据中识别这一点。其次,在实践中,我们只能访问感兴趣的语言特性的噪声代理-例如,从分类器和词典的预测。我们提出了一个估计这个设置,并证明其偏差是有界的,当我们执行的文本调整。基于这些结果,我们介绍了TextCause,估计语言属性的因果效应的算法。该方法利用(1)远程监督来提高噪声代理的质量,以及(2)预训练的语言模型(BERT)来调整文本。我们表明,所提出的方法优于相关方法估计亚马逊评论情绪的影响时,半模拟销售数字。最后,我们提出了一个应用案例研究调查投诉礼貌对官僚响应时间的影响。

英文摘要:

We consider the problem of using observational data to estimate the causal effects of linguistic properties. For example, does writing a complaint politely lead to a faster response time? How much will a positive product review increase sales? This paper addresses two technical challenges related to the problem before developing a practical method. First, we formalize the causal quantity of interest as the effect of a writer's intent, and establish the assumptions necessary to identify this from observational data. Second, in practice, we only have access to noisy proxies for the linguistic properties of interest -- e.g., predictions from classifiers and lexicons. We propose an estimator for this setting and prove that its bias is bounded when we perform an adjustment for the text. Based on these results, we introduce TextCause, an algorithm for estimating causal effects of linguistic properties. The method leverages (1) distant supervision to improve the quality of noisy proxies, and (2) a pre-trained language model (BERT) to adjust for the text. We show that the proposed method outperforms related approaches when estimating the effect of Amazon review sentiment on semi-simulated sales figures. Finally, we present an applied case study investigating the effects of complaint politeness on bureaucratic response times.

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。

↑