Extraction of script knowledge from texts. Part I. The task and the review of the state of the art

This paper discusses the importance of automatic extraction of script knowledge for natural language understanding. We discuss theoretical approaches to the description of text structure: story grammars, scripts, frames and narrative schemas. We provide a list of research fields where automatic script knowledge extraction can be applied to achieve better precision and recall (e.g. automatic summarization, information extraction, coreference resolution, etc.). The article also presents popular approaches to the automatic extraction of script knowledge and methods for evaluation of such approaches. Besides, we present a list of datasets that can be used to train and test new models.