arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1711.03230cs.CL
收藏

阅读理解任务中多轮推理策略的实证分析

An Empirical Analysis of Multiple-Turn Reasoning Strategies in Reading Comprehension Tasks

Microsoft Research(微软研究院);Johns Hopkins University(约翰斯·霍普金斯大学)

Yelong Shen, Xiaodong Liu, Kevin Duh, Jianfeng Gao

摘要:

阅读理解(RC)是一项具有挑战性的任务,需要跨句子的信息合成和多轮推理。使用最先进的RC模型,我们实证研究了SQuAD和MS MARCO数据集上单轮和多轮推理的性能。RC模型是一个具有迭代注意力的端到端神经网络,并使用强化学习来动态控制圈数。我们发现,多轮推理优于单轮推理的所有问题和答案类型;此外,我们观察到,使一个灵活的回合数一般提高了固定的多轮策略。%,并且对于具有冗长描述性答案的问题特别有益。我们在这两个数据集上取得了与最先进技术相竞争的结果。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑