arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1903.03033cs.CL
收藏

多项选择阅读理解的选项比较网络

Option Comparison Network for Multiple-choice Reading Comprehension

Pattern Recognition Center, WeChat AI, Tencent Inc(腾讯公司微信AI部模式识别中心);Tencent Inc(腾讯公司);WeChat AI(微信AI);Pattern Recognition Center(模式识别中心);China(中国)

Qiu Ran, Peng Li, Weiwei Hu, Jie Zhou

摘要:

多项选择阅读理解(MCRC)是从给定的问题和文章的多个选项中选择正确答案的任务。现有的MCRC模型通常要么独立读取每个选项,要么在比较它们之前为每个选项计算固定长度的表示。然而,人类通常在详细阅读文章之前在多粒度级别上比较选项,以使推理更有效。模仿人类,我们提出了一个选项比较网络(OCN)的MCRC比较选项在词级,以更好地确定它们的相关性,以帮助推理。特别地,每个选项都被编码成一个向量序列,使用一个skimmer,以保留尽可能细粒度的信息。利用注意力机制逐向量地比较这些序列,以识别选项之间更微妙的相关性,这对于推理可能是有价值的。在人类英语考试MCRC数据集RACE上的实验结果表明,我们的模型明显优于现有的方法。此外,它也是第一个在整个数据集上超越Amazon Mechanical Turker性能的模型。

英文摘要:

Multiple-choice reading comprehension (MCRC) is the task of selecting the correct answer from multiple options given a question and an article. Existing MCRC models typically either read each option independently or compute a fixed-length representation for each option before comparing them. However, humans typically compare the options at multiple-granularity level before reading the article in detail to make reasoning more efficient. Mimicking humans, we propose an option comparison network (OCN) for MCRC which compares options at word-level to better identify their correlations to help reasoning. Specially, each option is encoded into a vector sequence using a skimmer to retain fine-grained information as much as possible. An attention mechanism is leveraged to compare these sequences vector-by-vector to identify more subtle correlations between options, which is potentially valuable for reasoning. Experimental results on the human English exam MCRC dataset RACE show that our model outperforms existing methods significantly. Moreover, it is also the first model that surpasses Amazon Mechanical Turker performance on the whole dataset.

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。

↑