Deep Understanding Based Multi-Document Machine Reading Comprehension

Most existing multi-document machine reading comprehension models mainly focus on understanding the interactions between input question and documents, but ignore following two kinds of understandings. First, to understand semantic meaning words in documents from perspective each other. Second, supporting cues for a correct answer intra-document inter-documents. Ignoring these important understandings would make oversee some information that may be helpful inding answers. To overcome this deiciency, we propose deep based model comprehension. It has three cascaded modules which are designed accurate words, answer. We evaluate our large scale benchmark datasets, namely TriviaQA Web DuReader. Extensive experiments show achieves state-of-the-art results both datasets.