[paper] DuReader

DuReader: a Chinese Machine Reading Comprehension Dataset from Real-world Applications

Paper: https://arxiv.org/abs/1711.05073

Page: http://ai.baidu.com/broad/subordinate?dataset=dureader

Code: https://github.com/baidu/DuReader/

DuReader，一个新的大型开放中文机器阅读理解数据集。

DuReader 与以前的 MRC 数据集相比有三个优势：

1. 数据来源：问题和文档均基于百度搜索和百度知道; 答案是手动生成的。

2. 问题类型：它为更多的问题类型提供了丰富的注释，特别是是非类和观点类问题。

3. 规模：包含 200K 个问题，420K 个答案和 1M 个文档; 是目前最大的中文 MRC 数据集。

This paper introduces DuReader, a new large-scale, open-domain Chinese machine reading comprehension (MRC) dataset, designed to address real-world MRC.

DuReader has three advantages over previous MRC datasets:

1. data sources: questions and documents are based on Baidu Search and Baidu Zhidao; answers are manually generated.

2. question types: it provides rich annotations for more question types, especially yes-no and opinion questions, that leaves more opportunity for the research community.

3. scale: it contains 200K questions, 420K answers and 1M documents; it is the largest Chinese MRC dataset so far.

Introduction

Dataset	Lang	Que.	Docs	Source of Que.	Source of Docs	Answer Type
CNN/DM	EN	1.4M	300K	Synthetic cloze	News	Fill in entity
HLF-RC	ZH	100K	28K	Synthetic cloze	Fairy/News	Fill in word
CBT	EN	688K	108	Synthetic cloze	Children’s books	Multi. choices
RACE	EN	870K	50K	English exam	English exam	Multi. choices
MCTest	EN	2K	500	Crowd-sourced	Fictional stories	Multi. choices
NewsQA	EN	100K	10K	Crowd-sourced	CNN	Span of words
SQuAD	EN	100K	536	Crowd-sourced	Wiki.	Span of words
SearchQA	EN	140K	6.9M	QA site	Web doc.	Span of words
TrivaQA	EN	40K	660K	Trivia websites	Wiki./Web doc.	Span/substring of words
NarrativeQA	EN	46K	1.5K	Crowd-sourced	Book&movie	Manual summary
MS-MARCO	EN	100K	200K	User logs	Web doc.	Manual summary
DuReader	ZH	200k	1M	User logs	Web doc./CQA	Manual summary

表 1: 机器阅读理解数据集对比

Pilot Study

Examples	Fact	Opinion
Entity	iphone哪天发布	2017最好看的十部电影
-	On which day will iphone be released	Top 10 movies of 2017
Description	消防车为什么是红的	丰田卡罗拉怎么样
-	Why are firetrucks red	How is Toyota Carola
YesNo	39.5度算高烧吗	学围棋能开发智力吗
-	Is 39.5 degree a high fever	Does learning to play go improve intelligence

表 2: 中文六类问题的例子

Scaling up from the Pilot to DuReader

Data Collection and Annotation

Data Collection

DuReader 的样本可用四元组表示： [Math Processing Error] { q , t , D , A } ，其中 [Math Processing Error] q 是问题，[Math Processing Error] t 是问题类型，[Math Processing Error] D 是相关文档集合，[Math Processing Error] A 是由人类标注产生的答案集合。

The DuReader is a sequence of 4-tuples: [Math Processing Error] { q , t , D , A } , where [Math Processing Error] q is a question, [Math Processing Error] t is a question type, [Math Processing Error] D is a set of relevant documents, and [Math Processing Error] A is an answer set produced by human annotators.

Question Type Annotation

Answer Annotation

众包

Crowd-sourcing

Quality Control

Training, Development and Test Sets

数量	训练集	开发集	测试集
问题	181K	10K	10K
文档	855K	45K	46K
答案	376K	20K	21K

The training, development and test sets consist of 181K, 10K and 10K questions, 855K, 45K and 46K documents, 376K, 20K and 21K answers, respectively.

DuReader is (Relatively) Challenging

challenges:

1. The number of answers.

   

   图 1. 答案数量分布

2. The edit distance.

   人类生成的答案和源文档之间的差异很大。

   the difference between the human generated answers and the source documents is large.

3. The document length.

   问题平均 4.8 词，答案平均 69.6 词，文档平均 396 词。

   In DuReader, questions tend to be short (4.8 words on average) compared to answers (69.6 words), and answers tend to be short compared to documents (396 words on average).

Experiments

Baseline Systems

1. 从每个文件中选择一个最相关的段落

2. 在选定的段落中应用最先进的 MRC 模型

our designed systems have two steps:

1. select one most related paragraph from each document

2. apply the state-of-the-art MRC models on the selected paragraphs

Paragraph Selection

在训练阶段，我们从文档中选择与人类生成答案重叠最大的段落作为最相关段落。

In training stage, we select one paragraph from a document as the most relevant one, if the paragraph has the largest overlap with human generated answer.

在测试阶段，由于我们没有人类生成答案，我们选择与问题重叠最大的段落作为最相关段落。

In testing stage, since we have no human generated answer, we select the most relevant paragraph that has the largest overlap with the corresponding question.

Answer Span Selection

• Match-LSTM

  要在段落中找到答案，它会按顺序遍历段落，并动态地将注意力加权问题表示与段落的每个标记进行匹配。

  最后，使用答案指针层来查找段落中的答案范围。

  To find an answer in a paragraph, it goes through the paragraph sequentially and dynamically aggregates the matching of an attention-weighted question representation to each token of the paragraph.

  Finally, an answer pointer layer is used to find an answer span in the paragraph.

• BiDAF

  它使用上下文对问题的关注和问题对上下文的关注，以突出问题和上下文中的重要部分。

  之后，使用注意流层来融合所有有用的信息，以获得每个位置的向量表示。

  It uses both context-to-question attention and question-to-context attention in order to highlight the important parts in both question and context.

  After that, the so-called attention flow layer is used to fuse all useful information in order to get a vector representation for each position.

Results and Analysis

评价方法：BLEU-4, Rouge-L

We evaluate the reading comprehension task via character-level BLEU-4 and Rouge-L.

Systems	Baidu Search	-	Baidu Zhidao	-	All	-
-	BLEU-4%	Rouge-L%	BLEU-4%	Rouge-L%	BLEU-4%	Rouge-L%
Selected Paragraph	15.8	22.6	16.5	38.3	16.4	30.2
Match-LSTM	23.1	31.2	42.5	48.0	31.9	39.2
BiDAF	23.1	31.1	42.2	47.5	31.8	39.0
Human	55.1	54.4	57.1	60.7	56.1	57.4

Table 6: Performance of typical MRC systems on the DuReader.

Question type	Description	-	Entity	-	YesNo	-
-	BLEU-4%	Rouge-L%	BLEU-4%	Rouge-L%	BLEU-4%	Rouge-L%
Match-LSTM	32.8	40.0	29.5	38.5	5.9	7.2
BiDAF	32.6	39.7	29.8	38.4	5.5	7.5
Human	58.1	58.0	44.6	52.0	56.2	57.4

Table 8: Performance on various question types.

Opinion-aware Evaluation

Question type	Fact	-	Opinion	-
-	BLEU-4%	Rouge-L%	BLEU-4%	Rouge-L%
Opinion-unaware	6.3	8.3	5.0	7.1
Opinion-aware	12.0	13.9	8.0	8.9

Table 9: Performance of opinion-aware model on YesNo questions.

Discussion

Conclusion

提出了 DuReader 数据集，提供了几个 baseline。