# V-NET 多文档阅读理解任务思路论文（上）

## 文章内容

Abstract
Machine reading comprehension (MRC) on real web data usually requires the machine to answer a question by analyzing multiple passages retrieved by search engine. Compared with MRC on a single passage, multi-passage MRC is more challenging, since we are likely to get multiple confusing answer candidates from different passages. To address this problem, we propose an end-to-end neural model that enables those answer candidates from different passages to verify each other based on their content representations. Specially, we jointly train three modules that can predict the final answer based on three factors: the answer boundary, the answer content and the cross-passage answer verification. The experimental results show that our method outperforms the baseline by a large margin and achieves the state-of-the-art performance on the English MS-MARCO dataset and the Chinese DuReader dataset, both of which are designed for MRC in real-world settings.

### V-NET 亮点

• 问题-文章对/Q-P pair
• 指针网络/Pointer Network

### Q-P Pair

#### 注意点

$u_{t}^{Q}=BiLSTM(u_{t-1}^{Q},[e_{t}^{Q},c_{t}^{Q}]) \ u_{t}^{P_{i}}=BiLSTM(u_{t-1}^{P_{i}},[e_{t}^{P_{i}},c_{t}^{P_{i}}])$

$S_{t,k}=u_{t}^{QT} \cdot u_{k}^{P_{i}}$

$v^{P_{n}}=BiLSTM_{M}(v_{t-1}^{P_{i}},\tilde{ u }_{t}^{P_{i}})$

### 答案范围确认 - 指针网络

$-\frac{1}{N}\sum_{i=1}^{N}(\log a_{y_{i}^{1}}^{1}+\log a_{y_{i}^{2}}^{2})$其中 $N$ 是在数据中的样本个数，而 $y_{i}^{1}$ 和 $y_{i}^{2}$ 分别对应了答案区域的开头与结尾。然后我们就可以得到相应的答案。

## 参考文献

[1] Min Joon Seo, Aniruddha Kembhavi, Ali Farhadi, and Hannaneh Hajishirzi. 2016. Bidirectional attention ﬂow for machine comprehension. arXiv preprint arXiv:1611.01603.
[2] Wang S, Jiang J. Machine comprehension using match-lstm and answer pointer[J]. arXiv preprint arXiv:1608.07905, 2016.

# Bon