With current big data age, the magnitude of information become absolutely large. Consequently, how to extract required information from massive website resource more quickly becomes more and more important. The target of information extraction is to solve such problem, and it is the significant part of natural language processing. Information extraction is to transform information from unstructured text to structured knowledge, which can make information in the text more accessible for further processing.
The development of information extraction has about the history of 50 years. In the early stage, most works are completed by linguistic experts, which hope to design adequate linguistic templates to cover all situations of language expression. However, with the present enormous amount of information, such strategy must be awkward. So, more and more machine learning technologies have been exploited. My presentation is to briefly introduce the development of information extraction and the representative approaches at different stages.

Pengda Qin received the bachelor degree from Beijing University of Post and Telecommunications in 2014, and he is now pursuing the Ph.D. degree in school of information and communication engineering at Beijing University of Post and Telecommunications. His research interests include machine learning, deep learning and information extraction etc. His paper “A Novel Negative Sampling based on TFIDF for learning word representation” has been received by the journal of Neurocomputing.