Automatic Semantic Annotation for Abstracts of Scientific Discourses.
Lan Huang, Jinchao Zhu, Chao Yang, Hao Xu · 2015
The abstract of scientific papers has strong semantic structure, which contains abundant meaningful information, such as the background, research problem, solution, and result. Marking it out can help the computer understand and use this underlying information, which can provide great help for searching and scanning papers. In order to annotate the semantics of the paper automatically, we modeled the rhetorical structure of an abstract by linguistic clues and position information.