An Empirical Analysis of Multiple-Turn Reasoning Strategies in Reading Comprehension Tasks
This work addresses reading comprehension tasks for natural language processing applications, but it is incremental as it builds on existing state-of-the-art models with empirical analysis.
The paper tackled the problem of improving reading comprehension by comparing single-turn and multiple-turn reasoning strategies, finding that multiple-turn reasoning outperforms single-turn reasoning across all question types and that a flexible number of turns further enhances performance, achieving competitive results on SQuAD and MS MARCO datasets.
Reading comprehension (RC) is a challenging task that requires synthesis of information across sentences and multiple turns of reasoning. Using a state-of-the-art RC model, we empirically investigate the performance of single-turn and multiple-turn reasoning on the SQuAD and MS MARCO datasets. The RC model is an end-to-end neural network with iterative attention, and uses reinforcement learning to dynamically control the number of turns. We find that multiple-turn reasoning outperforms single-turn reasoning for all question and answer types; further, we observe that enabling a flexible number of turns generally improves upon a fixed multiple-turn strategy. %across all question types, and is particularly beneficial to questions with lengthy, descriptive answers. We achieve results competitive to the state-of-the-art on these two datasets.