paper

English phrases for writing

1. 定义术语 We use the term "xx" to refer to a context passage paired with an unanswerable question.
2. 给baseline起名字 we refer to these as xxx
3. In extractive reading comprehension datasets, a system must extract the correct answer to a question from a context document or paragraph.
4. Sentence selection datasets test whether a system acan rank sentences that answer a question higher than sentences that do not.
5. sentence selection(QASENT, WikiQA) 和 multiple choice （MCTest, RACE）是不同的任务
6. are not able to doing 不能做
7. tend to do sth 可能做
8. Consequently, we first given a brief introduction to ..., and then investigate .
9. be capable of jointly doing sth 能做某事
10. considerable performance gains 比较大的性能提升
11. The statistics of the xxx dataset is shown in the Table x.
12. We use the term xxx to refer to xxx with xxx.
13. IN extractive reading comprehension datasets, a system must extract the correct answer to a question from a context document or paragraph.
14. user-facing system
15. Recognizeing textual entailment requires systems to decide whether a hypothesis is entailed by, contradicted by, or neural with respect to a premise.
16. character-derived embeddings
17. restrict their applicability in many domains that suffers from a dearth of ... 受限制
18. provide a valuable alternatives to doning sth. 提供一种可选方案
19. provide a significant performance boost. 提升一个显著的性能提升
20. compelling evidence for 有说服力的证据
21. doing sth is challenging for two reason. First, xxx. Second, xxx. 做某事是有挑战的因为两个原因，其一。其二。
22. This model choice provides us with more memmory for handling .. 模型结构的选择使得我们有更多空间去处理
23. compared to alternatives like ... ， resulting in
24. Our training proedure consists of two stages. The first stage is ... . This is followed by a xx stage, where we adapt the model to a discriminative task with labels data
25. must posses the ability to determine whether ..
26. if and only if the associated passage contains the answer
27. strong baseline system that use pieline and threshold-based approaches
28. Machine comprehension systems mimic the process of reading comprehension by ansewring questions after understanding natural language text.
29. neural model for answer span extraction
30. Although this does allow us to obtain bidirectional pre-trained model, threr are two main downsides to such an approch.
31. semantically equivalent 语义等价
32. Given a pair of sentences, the goal is to predict whether the second sentence is an entailment, contradication, or neutral with respect to the first one.
33. is more concerned with doing /put more emphasis on N 更加关注
34. Types of machine learning 机器学习的累心
35. almost all 几乎所有
36. preliminaries 前言 background 背景， 作为第二节
37. in principle 原则上，大体上
38. is free of 免疫 is free of adversarial attck
39. is upper bounded by
40. For ease of exposition 为了说明简单