Wikidata, growing linked data which is full potential to provide reliable and powerful service, bases on the knowledge of Wikipedia. Although multiple projects required information on Wikidata, the property are not as sufficient as Wikipedia, even some property values on Wikidata are not right. How to extract knowledge from Wikipedia automatically and add it to Wikidata is a worthy of attention.
In this article, we study an effective method for extracting Wikidata property values. Specifically, we take the Wikipedia text as input 1, take the property question or property definition as input 2, and use the property value as the output answer. Use ALBERT for machine understanding, predict the starting and ending positions as output, and compare the two groups of predicted answers.