Argumentative structure of scientific abstracts - V0.1

In order to explore the possibility of leveraging discourse information for the identiffication of argumentative components and relations we add a new annotation layer to a subset of the Discourse Dependency TreeBank for Scientiffic Abstracts (SciDTB). [1] We introduce a ffine-grained annotation schema aimed at capturing information that accounts for the specifficities of the scientiffic discourse, including the type of evidence that is offered to support a statement (e.g., background information, experimental data or interpretation of results). [1] Yang, A., Li, S.: SciDTB: Discourse dependency TreeBank for scientiffc abstracts. In: Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (ACL 2018) (Volume 2: Short Papers). pp. 444{449. Association for Computational Linguistics, Melbourne, Australia (Jul 2018)

为探索利用语篇信息识别论辩性成分与论辩关系的可能性，我们为科学摘要语篇依存树库（Discourse Dependency TreeBank for Scientific Abstracts，简称SciDTB）的子集新增了一层标注。[1] 我们提出了细粒度标注方案，旨在捕捉契合科学语篇特性的信息，其中包括用于支撑学术陈述的各类证据类型，例如背景信息、实验数据或结果解读。[1] Yang, A., Li, S.: SciDTB: 科学摘要语篇依存树库。载于：第56届国际计算语言学协会年会（ACL 2018）论文集（第2卷：短文）。第444-449页。国际计算语言学协会，澳大利亚墨尔本（2018年7月）