Composing Answer from Multi-spans for Reading Comprehension
arXiv:2009.06141
Abstract
This paper presents a novel method to generate answers for non-extraction machine reading comprehension (MRC) tasks whose answers cannot be simply extracted as one span from the given passages. Using a pointer network-style extractive decoder for such type of MRC may result in unsatisfactory performance when the ground-truth answers are given by human annotators or highly re-paraphrased from parts of the passages. On the other hand, using generative decoder cannot well guarantee the resulted answers with well-formed syntax and semantics when encountering long sentences. Therefore, to alleviate the obvious drawbacks of both sides, we propose an answer making-up method from extracted multi-spans that are learned by our model as highly confident -gram candidates in the given passage. That is, the returned answers are composed of discontinuous multi-spans but not just one consecutive span in the given passages anymore. The proposed method is simple but effective: empirical experiments on MS MARCO show that the proposed method has a better performance on accurately generating long answers, and substantially outperforms two competitive typical one-span and Seq2Seq baseline decoders.
arXiv commentDue to the policy of our institute, with the agreement of all of the author, we decide to withdraw this paper
References in corpus (6)
- Sequence to Sequence Learning with Neural Networks
- Teaching Machines to Read and Comprehend
- Bidirectional Attention Flow for Machine Comprehension
- The Goldilocks Principle: Reading Children's Books with Explicit Memory Representations
- SearchQA: A New Q&A Dataset Augmented with Context from a Search Engine
- Machine Reading Comprehension: The Role of Contextualized Language Models and Beyond