Automatically Generating Expository Text by Modeling Wikipedia

点击翻译按钮获取中文摘要

Aiming to efficiently utilize descriptions for various terms, we proposed a method to automatically generate expository text like an article in Wikipedia, from a set of pages on the Web. Because viewpoints required for describing a term are different depending on the type of that term, such as animal and disease, we use Wikipedia articles to learn viewpoint-based patterns for term descriptions. Depending on the type of a target term, we extract required sentences from a search result and integrate them into an expository text.