9 In this paper , we propose a computer system that uses machine learning to automatically identify the structure of texts , enabling researchers to quickly and effectively process very large corpora .