论文写作套路之一如何写摘要

论文写作套路之一如何写摘要发布时间：2019-06-27 11:04:34

论文写作套路之一：如何写摘要

如果说标题（Title）是paper的眼，那么摘要（Abstract）就是paper的脸面。没有漂亮的脸面，读者或者审稿人没有继续往下看的好奇心和动力。因此，论文的摘要写作是极其重要的，是作者写作水平和研究能力的完美体现。
Title和Abstract结合起来相当于一篇具有独立性和自明性的短文，即不用再阅读全文就能获得论文中的必要信息。它是作者对该篇论文精华部分的高度概括，用简单、明确、易懂、精辟的语言反映作者的观点、主要内容、研究成果、独到的见解，分层次，条理性强。摘要写作强调四要素，即目的、方法、结果（或过程）、结论。摘要必须回答“研究什么”、“怎么研究”、“得到了什么结果”、“结果说明了什么”等。
SCI论文通常采用结构化的摘要，其基本组成部分是：
目的/问题：我们为什么要关注该问题？当前该问题存在哪些研究上的不足？如果研究的问题非常普遍，大家基本都了解，可简单地描述；
方法/过程：用什么方法通过什么步骤得到研究结果？
结果：取得了什么样的研究结果？
结论：通过分析结果，得到了什么结论？是否对第一点提到的研究不足有改进（质量提升、性能优化）？
依据所研究问题的特点，为了突显各个部分的重要性，以上4个要素在摘要里的权重或详略程度会不同。比如，为了突显研究问题的紧迫性或重要性，可以适当多描述目的/问题部分；而如果该问题比较普遍，就可以非常简单的一笔带过。同样的，在研究方法部分，如果旨在凸显研究方法的创新性，即可着重笔墨；而如果该研究方法（比如某种试验技术）很成熟，就可以只写出试验方法而不描述过程。
总之，好的摘要一定是主次分明且有逻辑。写作过程中，尤其要注意以下几点：
（1）摘要中应排除本学科领域已成为常识的内容；切忌把应在引言中出现的内容写入摘要；一般也不要对论文内容作诠释和评论（尤其是自我评价）。
（2）不得简单重复标题中已有的信息。
（3）尽量用第三人称（当然现在有些杂志也用第一人称）。
（4）除了实在无法变通以外，一般不用数学公式，不出现插图、插表。
（5）不用引文，除非该文献证实或否定了他人已出版的著作。
（6）缩略语、略称、代号，除了相邻专业的读者也能清楚理解的以外，在首次出现时必须加以说明。
（7）切忌不可将摘要、序、结论写的一样。
（8）摘要、引言、结论三者之间的关系：①“摘要”应有结论性的语言；②“引言”不应出现结论性的语言；③“结论”不应出现背景介绍、论文主要分析部分。
这里我选择8篇关于机器学习的论文摘要进行对比，分析他们是怎么写的。下面来看例子：
1）Convolutional Pose Machines（论文标题）
Pose Machines provide a sequential prediction framework for learning rich implicit spatial models.
本摘要没有过多的铺垫，直奔主题，第一句首先提到了姿态机。
In this work we show a systematic design for how convolutional networks can be incorporated into the pose machine framework for learning image features and image-dependent spatial models for the task of pose estimation.
然后直接说明本文的意图，即介绍如何设计卷积神经网络，将其包括到姿态机这个框架中来学习图像的特征和依赖图像的空间模型进而实现姿态估计。
The contribution of this paper is to implicitly model long-range dependencies between variables in structured prediction tasks such as articulated pose estimation.
其后直接说出本文的贡献：即隐式地对长范围的依赖进行建模。
We achieve this by designing a sequential architecture composed of convolutional networks that directly operate on belief maps from previous stages, producing increasingly refined estimates for part locations, without the need for explicit graphical model-style inference.
而后说出本文具体是怎么做的：通过设计包含CNN的网络结构，该网络结构能够在前一阶段的belief map的结果之上进行，这样可以逐渐地得到经过精化之后的身体部件的位置，这种方式不需要显式地进行图模型建模。
Our approach addresses the characteristic difficulty of vanishing gradients during training by providing a natural learning objective function that enforces intermediate supervision, thereby replenishing back-propagated gradients and conditioning the learning procedure. We demonstrate state-of-the-art performance and outperform competing methods on standard benchmarks including the MPII, LSP, and FLIC datasets.
最后给出了本文方法所强调要解决的问题，并且说明本方法NB，state of art performance。
点评：这种写法输入一般套路写法

首先简单铺垫；
本文的大体意图；
一句话说明贡献；
一句话说明具体是怎么做的

⑤最后给出本文所的方法所解决的困难是什么，本文方法在xxx数据集上取得了state of art结果
可以使用的语法结构：
你方法的大体意图
in this work wo show a systematic design for xxxx
介绍你的贡献
the contribution of this paper is to xxx
用于介绍你自己的方法
we achieve this by xxxx
说明你方法NB
our approach address xxx problem by xxxxx. We demonstrate state-of-the-art performance and outperform competing methods on standard benchmarks including the xxxx

2） End-to-End Learning of Deformable Mixture of Parts and Deep Convolutional Neural Networks for Human Pose Estimation
Recently, Deep Convolutional Neural Networks (DCNNs) have been applied to the task of human pose estimation, and have shown its potential of learning better feature representations and capturing contextual relationships.
首先介绍CNN用于人的姿态估计，并且能够捕获好的特征。
However, it is difficult to incorporate domain prior knowledge such as geometric relationships among body parts into DCNNs.
引出要解决的问题：难以包含领域先验知识，比如身体部件之间的几何关系到CNN中
In addition, training DCNN-based body part detectors without consideration of global body joint consistency introduces ambiguities, which increases the complexity of training.
进一步突出要解决的问题重要性：即如果不考虑身体部件之间的关系会导致引入歧义，增加模型训练复杂度
In this paper, we propose a novel end-to-end framework for human pose estimation that combines DCNNs with the expressive deformable mixture of parts.
接着提出本文的方法：端到端的框架，能够将CNN和可变部件模型结合起来
We explicitly incorporate domain prior knowledge into the framework, which greatly regularizes the learning process and enables the flexibility of our framework for loopy models or tree-structured models.
具体介绍本文方法如何做的：显式地将领域先验知识加入到框架，这样做能够对学习过程起到正则化的作用，保证框架的灵活性
The effectiveness of jointly learning a DCNN with a deformable mixture of parts model is evaluated through intensive experiments on several widely used benchmarks.
The proposed approach significantly improves the performance compared with state-of-the-art approaches, especially on benchmarks with challenging articulations.
说明方法的结果：在若干个benchmarks上NB，又是state of art…….
点评：一般套路写法

铺垫
引出要解决的问题，突触要解决的问题的重要性
一句话介绍你的方法
一句话介绍具体如何做的

⑤你的方法NB，在xxx数据集行state of art
善用连词：
一般用在引出你所要解决的问题上
However
用于进一步强调如果不解决这个问题会怎么样
In addition
用于介绍你自己的方法
in this paper we propose a novel end-to end framework for xxxx that xxxxxxxx
说明你方法NB
The effectiveness of xxxxx is evaluated througth intensive experiments on xxxx benchmarks, the proposed approach significantly improves the performance compared with state-of-the-art approaches

3）Human Pose Estimation with Iterative Error Feedback
Hierarchical feature extractors such as Convolutional Networks (ConvNets) have achieved impressive performance on a variety of classification tasks using purely feedforward processing.
铺垫：CNN能够提取层次特征
Feedforward architectures can learn rich representations of the input space but do not explicitly model dependencies in the output spaces, that are quite structured for tasks such as articulated human pose estimation or object segmentation.
引出要解决的问题：前馈结构可以学习到输入空间的丰富的表达但是不能显式地建模在输出空间的依赖关系。
Here we propose a framework that expands the expressive power of hierarchical feature extractors to encompass both input and output spaces, by introducing top-down feedback.
提出本文的方法：提出一种扩展层次特征提取器的表达能力的方法即将通过引入top-down的反馈将输入和输出空间都包括进来
Instead of directly predicting the outputs in one go, we use a self-correcting model that progressively changes an initial solution by feeding back error predictions, in a process we call Iterative Error Feedback (IEF).
进一步突出本文方法的区别：使用自矫正模型，通过将错误预测的无法反馈回去，渐进地改变初始解，本文称之为IEF
IEF shows excellent performance on the task of articulated pose estimation in the challenging MPII and LSP benchmarks, matching the state-of-the-art without requiring ground truth scale annotation.
说明本文方法NB：IEF NB
点评：一般套路写法

一句话铺垫
引出要解决的问题
提出本文方法
突出本文方法的区别
突出自己方法NB

可用的语法结构：
用于引出要解决的问题
xxxx is good but do not xxxxx
提出本文方法
Here we propose a framework that xxxx, by xxxxx
突出本文方法
Instead of xxx, we use xxxx
突出自己方法NB
Our method show excellent performance on the task of xxx in the challenging xxx and xxx benchmarks

4）Personalizing Human Video Pose Estimation
We propose a personalized ConvNet pose estimator that automatically adapts itself to the uniqueness of a person’s appearance to improve pose estimation in long videos.
没有铺垫，直接提出本文方法：提出了个性化的卷积神经网络姿态估计器，该方法能够自动地将其适配到某个人的外形上去，从而改善长时间段的视频中的人的姿态估计性能
We make the following contributions:
(i) we show that given a few high-precision pose annotations, e.g. from a generic ConvNet pose estimator, additional annotations can be generated throughout the video using a combination of image-based matching for temporally distant frames, and dense optical flow for temporally local frames;
(ii) we develop an occlusion aware self-evaluation model that is able to automatically select the high-quality and reject the erroneous additional annotations;
(iii) we demonstrate that these high-quality annotations can be used to fine-tune a ConvNet pose estimator and thereby personalize it to lock on to key discriminative features of the person’s appearance.
直接说出本文贡献：三个点
The outcome is a substantial improvement in the pose estimates for the target video using the personalized ConvNet compared to the original generic ConvNet.
本文方法NB
Our method outperforms the state of the art (including top ConvNet methods) by a large margin on two standard benchmarks, as well as on a new challenging YouTube video dataset. Furthermore, we show that training from the automatically generated annotations can be used to improve the performance of a generic ConvNet on other benchmarks.
本文方法还是NB
点评：直奔主题的写法套路

提出本文方法
给出本文贡献，逐个列出来
强调本文方法NB

可用的语法结构：
提出本文方法：
We propose a xxx that xxxx
给出本文贡献
We make the following contributions:
(i) we show that xxx
(ii) we develop a xxx
(iii) we demonstrate that xxx
本文方法NB
Our method outperforms the state of the art by a large margin on two standard benchmarks, as well as on a new challenging YouTube video dataset
5）Structured Feature Learning for Pose Estimation
In this paper, we propose a structured feature learning framework to reason the correlations among body joints at the feature level in human pose estimation.
提出本文方法：结构化特征学习框架用于在特征层次上推断关节点之间的关系
Different from existing approaches of modeling structures on score maps or predicted labels, feature maps preserve substantially richer descriptions of body joints.
强调本文方法的区别：现有方法都是对在score map或者预测类标上的结构进行建模，而feature map则保留了大量更加丰富的关节描述信息
The relationships between feature maps of joints are captured with the introduced geometrical transform kernels, which can be easily implemented with a convolution layer.
Features and their relationships are jointly learned in an end-to-end learning system. A bi-directional tree structured model is proposed, so that the feature channels at a body joint can well receive information from other joints.
本文方法具体怎么做
The proposed framework improves feature learning substantially. With very simple post processing, it reaches the best mean PCP on the LSP and FLIC datasets. Compared with the baseline of learning features at each joint separately with ConvNet, the mean PCP has been improved by 18% on FLIC. The code is released to the public.
本文方法NB
点评：直奔主题的写法套路

提出本文方法
强调本文方法的区别
本文方法具体怎么做
本文方法NB

可用的语法结构：
提出本文方法
In this paper, we propose a xxx learning framework to xxx
强调本文方法的区别
Different from existing approaches of xxx, sssss can xxxxx
本文方法NB
The proposed framework improves xxxxx substantially. With xxxxx, it reaches the best mean PCP on the xxx and sss datasets. Compared with the baseline of xxx the mean PCP has been improved by 18% on FLIC.

6）Chained Predictions Using Convolutional Neural Networks
In this paper, we present an adaptation of the sequence-tosequence model for structured output prediction in vision tasks.
提出本文方法
In this model the output variables for a given input are predicted sequentially using neural networks. The prediction for each output variable depends not only on the input but also on the previously predicted output variables. The model is applied to spatial localization tasks and uses convolutional neural networks (CNNs) for processing input images and a multi-scale deconvolutional architecture for making spatial predictions at each time step. We explore the impact of weight sharing with a recurrent connection matrix between consecutive predictions, and compare it to a formulation where these weights are not tied. Untied weights are particularly suited for problems with a fixed sized structure, where different classes of output are predicted in different steps.
本文方法具体怎么做
We show that chained predictions achieve top performing results on human pose estimation from single images and videos.
本文方法NB

点评：直奔主题的写法套路

提出本文方法
本文方法具体怎么做
本文方法NB

可用的语法结构：
提出本文方法
In this paper, we present an adaptation of xxx model for xxx in vision tasks.
本文方法NB
We show that our method achieves top performing results on human pose estimation from single images and videos.

7）Stacked Hourglass Networks for Human Pose Estimation
This work introduces a novel convolutional network architecture for the task of human pose estimation.
提出本文方法
Features are processed across all scales and consolidated to best capture the various spatial relationships associated with the body.
We show how repeated bottom-up, top-down processing used in conjunction with intermediate supervision is critical to improving the performance of the network. We refer to the architecture as a \stacked hourglass” network based on the successive steps of pooling and upsampling that are done to produce a final set of predictions.
本文方法具体怎么做
State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.
本文方法NB
点评：直奔主题的写法套路

提出本文方法
本文方法具体怎么做

③ 本文方法NB

8）Multi-Person Pose Estimation with Local Joint-to-Person Associations
Despite of the recent success of neural networks for human pose estimation, current approaches are limited to pose estimation of a single person and cannot handle humans in groups or crowds.
引出要解决的问题
In this work, we propose a method that estimates the poses of multiple persons in an image in which a person can be occluded by another person or might be truncated.
提出本文方法
To this end, we consider multiperson pose estimation as a joint-to-person association problem.
We construct a fully connected graph from a set of detected joint candidates in an image and resolve the joint-to-person association and outlier detection using integer linear programming. Since solving joint-to-person association jointly for all persons in an image is an NP-hard problem and even approximations are expensive, we solve the problem locally for each person.
具体介绍本文方法
On the challenging MPII Human Pose Dataset for multiple persons, our approach achieves the accuracy of a state-of-the-art method, but it is 6,000 to 19,000 times faster.
突出自己方法NB
点评：一般套路写法，没有铺垫

引出要解决的问题
提出本文方法
具体介绍本文方法
突出自己方法NB

可用的语法结构：
引出要解决的问题
Despite of the recent success of xxx, current approaches are limited to xxx.
突出自己方法NB
On the challenging xxx Dataset for sss, our approach achieves the accuracy of a state-of-the-art method

套路总结
（1）铺垫，这个因研究方向而异（用一句话或者一个短语来介绍研究目的）
（2）引出所要研究的问题（指出现有方法的缺陷）
However, it is difficult to xxxx
xxxx is good but do not xxxxx
Despite of the recent success of xxx, current approaches are limited to xxx.
（2.1）进一步强调如果不解决这个问题会怎么样（这段可以省略）
In addition, training DCNN-based body part detectors without consideration of global body joint consistency introduces ambiguities, which increases the complexity of training.
（3）介绍你自己的方法（或者贡献）
the contribution of this paper is to xxx
in this work we show a systematic design for xxxx
in this paper we propose a novel end-to end framework for xxxx that xxxxxxxx
In this paper, we propose a xxx framework to xxx
In this paper, we present an adaptation of xxx model for xxx in vision tasks.
Here we propose a framework that xxxx, by xxxxx
We propose a xxxx that can do xxxx
We achieve this by xxxx
We make the following contributions:
(i) we show that xxx
(ii) we develop a xxx
(iii) we demonstrate that xxx
（3.1）突出你的方法的区别（这段可以省略）
Instead of xxx, we use xxxx
Different from existing approaches of modeling ssss, xxx can aaaaaa
（4）说明你方法NB（两三句话来介绍一下细节, 主要结果及结论等）
Our approach address xxx problem by xxxxx. We demonstrate state-of-the-art performance and outperform competing methods on standard benchmarks including the xxxx
Our method show excellent performance on the task of xxx in the challenging xxx and xxx benchmarks
Our method outperforms the state of the art by a large margin on two standard benchmarks, as well as on a new challenging YouTube video dataset
The proposed framework improves xxxxx substantially. With xxxxx, it reaches the best mean PCP on the xxx and sss datasets. Compared with the baseline of xxx the mean PCP has been improved by 18% on FLIC.
The effectiveness of xxxxx is evaluated througth intensive experiments on xxxx benchmarks, the proposed approach significantly improves the performance compared with state-of-the-art approaches
We show that our method achieves top performing results on human pose estimation from single images and videos.
State-of-the-art results are achieved on the FLIC and MPII benchmarks outcompeting all recent methods.
On the challenging xxx Dataset for sss, our approach achieves the accuracy of a state-of-the-art method

再看一个经典的摘要-关于机器学习的例子（对比一下前面的套路想一想）：
“Mixtures of Gaussians are among the most fundamental and widely used statistical models. Current techniques for learning such mixtures from data are local search heuristics with weak performance guarantees. We present the first provably correct algorithm for learning a mixture of Gaussians. The algorithm is very simple and returns the true centers of the Gaussians to within the precision specified by the user, with high probability. It runs in time only linear in the dimension of the data and polynomial in the number of Gaussians.”
– Sanjoy Dasgupta in “Learning Mixtures of Gaussians”
“混合高斯模型是最基本、应用最广泛的统计模型之一。目前从数据中学习这种混合模型的方法是依靠局部搜索，无法保证效果。我们提出了第一个可证明的学习混合高斯模型的算法。该算法非常简单, 能在用户指定的精度范围内大概率地返回高斯模型的真实中心。算法能在数据维度的线性时间内，高斯模型数的多项式时间内返回结果”