End-to-End Answer Chunk Extraction and Ranking for Reading Comprehension

Yang Yu, Wei Zhang, Bowen Zhou, Kazi Hasan, Mo Yu, Bing Xiang

Nov 04, 2016 (modified: Jan 09, 2017) ICLR 2017 conference submission readers: everyone
  • Abstract: This paper proposes dynamic chunk reader (DCR), an end-to-end neural reading comprehension (RC) model that is able to extract and rank a set of answer candidates from a given document to answer questions. DCR is able to predict answers of variable lengths, whereas previous neural RC models primarily focused on predicting single tokens or entities. DCR encodes a document and an input question with recurrent neural networks, and then applies a word-by-word attention mechanism to acquire question-aware representations for the document, followed by the generation of chunk representations and a ranking module to propose the top-ranked chunk as the answer. Experimental results show that DCR could achieve a 66.3% Exact match and 74.7% F1 score on the Stanford Question Answering Dataset.
  • Conflicts: ibm.com
  • Keywords: Natural language processing, Deep learning, Supervised Learning