Open information extraction. Part I. The task and the review of the state of the art

The paper discusses the task of open information extraction from natural language texts. Open information extraction – is rather new approach to solving tasks of information extraction that do not specify structure and semantics of the information to be extracted. This approach is domain independent and does not require big annotated corpora. We present the formulation of the problem and review the state of the art related to extraction of entities and semantic relations from texts including methods of information extraction based on semi-supervised and unsupervised learning. We present the future directions of research of methods for relation extraction based on unsupervised learning.