End-to-End Answer Chunk Extraction and Ranking for Reading Comprehension
This addresses a limitation in neural reading comprehension models that primarily predict single tokens or entities, enabling more flexible answer extraction for tasks like question answering.
The paper tackles the problem of extracting variable-length answer chunks in reading comprehension, proposing the dynamic chunk reader (DCR) model that achieves state-of-the-art exact match and F1 scores on the SQuAD dataset.
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 achieves state-of-the-art exact match and F1 scores on the SQuAD dataset.