文章基本信息

标题：Belief-based Generation of Argumentative Claims
本地全文：下载
作者：Milad Alshomary ; Wei-Fan Chen ; Timon Gurcke 等
期刊名称：Conference on European Chapter of the Association for Computational Linguistics (EACL)
出版年度：2021
卷号：2021
页码：224-233
DOI：10.18653/v1/2021.eacl-main.17
语种：English
出版社：ACL Anthology
摘要：In this work, we argue that augmenting argument generation technology with the ability to encode beliefs is of twofold. First, it gives more control on the generated arguments leading to better reach for audience. Second, it is one way of modeling the human process of synthesizing arguments. Therefore, we propose the task of belief-based claim generation, and study the research question of how to model and encode a user’s beliefs into a generated argumentative text. To this end, we model users’ beliefs via their stances on big issues, and extend state of the art text generation models with extra input reflecting user’s beliefs. Through an automatic evaluation we show empirical evidence of the applicability to encode beliefs into argumentative text. In our manual evaluation, we highlight that the low effectiveness of our approach stems from the noise produced by the automatic collection of bag-of-words, which was mitigated by removing this noise. The finding of this paper lays the ground work to further investigate the role of beliefs in generating better reaching arguments.