A method of generating a question data set from a knowledge base including a plurality of statements includes generating at least one question template based on a structure of selected statements of the plurality of statements, generating a seed question for each selected statement based on the at least one question template with a processor, generating at least one first extension question with a search engine by processing each of the seed questions through the search engine, and storing at least one of the at least one first extension questions and the seed questions in a memory as the question data set.