Convolutional Spatial Attention Model for Reading Comprehension with Multiple-Choice Questions
This work addresses reading comprehension for AI systems, offering a novel method that improves performance on specific benchmarks, though it is incremental in nature.
The paper tackles machine reading comprehension with multiple-choice questions by proposing a Convolutional Spatial Attention model that extracts mutual information among passage, question, and candidates, achieving substantial improvements over state-of-the-art systems on RACE and SemEval-2018 Task11 datasets.
Machine Reading Comprehension (MRC) with multiple-choice questions requires the machine to read given passage and select the correct answer among several candidates. In this paper, we propose a novel approach called Convolutional Spatial Attention (CSA) model which can better handle the MRC with multiple-choice questions. The proposed model could fully extract the mutual information among the passage, question, and the candidates, to form the enriched representations. Furthermore, to merge various attention results, we propose to use convolutional operation to dynamically summarize the attention values within the different size of regions. Experimental results show that the proposed model could give substantial improvements over various state-of-the-art systems on both RACE and SemEval-2018 Task11 datasets.