Skip to content

M
mtbookv2
Commit ab602d82 by 曹润柘
Options

update

parent bf50e681
正在显示 包含 2497 行增加 和 0 行删除
ChapterAppend/chapterappend.tex 0 → 100644
% !Mode:: "TeX:UTF-8"
% !TEX encoding = UTF-8 Unicode
%----------------------------------------------------------------------------------------
% 机器翻译：统计建模与深度学习方法
% Machine Translation: Statistical Modeling and Deep Learning Methods
%
% Copyright 2020
% 肖桐(xiaotong@mail.neu.edu.cn) 朱靖波 (zhujingbo@mail.neu.edu.cn)
%----------------------------------------------------------------------------------------
%----------------------------------------------------------------------------------------
% CONFIGURATIONS
%----------------------------------------------------------------------------------------
\part{附录}
\renewcommand\figurename{图}%将figure改为图
\renewcommand\tablename{表}%将figure改为图
\chapterimage{../Figures/fig-NEU-1.jpg} % Chapter heading image
%----------------------------------------------------------------------------------------
% CHAPTER APPENDIX A
%----------------------------------------------------------------------------------------
\begin{appendices}
\chapter{附录A}
\label{appendix-A}
\parinterval 在构建机器翻译系统的过程中，数据是必不可少的，尤其是现在主流的神经机器翻译系统，系统的性能往往受限于语料库规模和质量。所幸的是，随着语料库语言学的发展，一些主流语种的相关语料资源已经十分丰富。
\parinterval 为了方便读者进行相关研究，我们汇总了几个常用的基准数据集，这些数据集已经在机器翻译领域中被广泛使用，有很多之前的相关工作可以进行复现和对比。同时，我们收集了一下常用的平行语料，方便读者进行一些探索。
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{基准数据集}
%----------------------------------------------
\begin{table}[htp]{
\footnotesize
\begin{center}
\caption{基准数据集}
\label{tab:Reference-data-set}
\begin{tabular}{p{1.6cm} | p{1.2cm} p{1.6cm} p{2.6cm} p{3.9cm}}
{任务} & {语种} &{领域} &{描述} &{数据集地址} \\
\hline
\rule{0pt}{15pt}WMT & En Zh& 新闻、医学 & 以英语为核心的多& {http://www.statmt.org/wmt19/} \\
& De Ru等 & 、翻译 & 语种机器翻译数据 & \\
& & & 集，涉及多种任务 & \\
\rule{0pt}{15pt}IWSLT & En De Fr & 口语翻译 & 文本翻译数据集来 & {https://wit3.fbk.eu/} \\
& Cs Zh等 & &自TED演讲，数 & \\
& & & 据规模较小 & \\
\rule{0pt}{15pt}NIST & Zh-En等 & 新闻翻译 & 评测集包括4句参 & {https://www.ldc.upenn.edu/coll} \\
& Cs Zh等 & & 考译文，质量较高 & aborations/evaluations/nist \\
\end{tabular}
\end{center}
}\end{table}
%----------------------------------------------
%----------------------------------------------
\begin{table}[htp]{
\footnotesize
\begin{center}
\begin{tabular}{p{1.6cm} | p{1.2cm} p{1.6cm} p{2.6cm} p{3.9cm}}
\rule{0pt}{15pt}{任务} & {语种} &{领域} &{描述} &{数据集地址} \\
\hline
\rule{0pt}{15pt}TVsub & Zh-En & 字幕翻译 & 数据抽取自电视剧 & {https://github.com/longyuewan} \\
& & & 字幕，用于对话中 & gdcu/tvsub \\
& & & 长距离上下文研究 & \\
\rule{0pt}{15pt}Flickr30K & En-De & 多模态翻译 & 31783张图片，每 & {http://shannon.cs.illinois.edu/D} \\
& & & 张图片5个语句标 & enotationGraph/ \\
& & & 注 & \\
\rule{0pt}{15pt}Multi30K & En-De & 多模态翻译 & 31014张图片，每 & {http://www.statmt.org/wmt16/} \\
& En-Fr & & 张图片5个语句标 & multimodal-task.html \\
& & & 注 & \\
\rule{0pt}{15pt}IAPRTC-12 & En-De & 多模态翻译 & 20000张图片及对 & {https://www.imageclef.org} \\
& & & 应标注 & /photodata \\
\rule{0pt}{15pt}IKEA & En-De & 多模态翻译 & 3600张图片及对应 & {https://github.com/sampalomad} \\
& En-Fr & & 标注 & /IKEA-Dataset.git \\
\end{tabular}
\end{center}
}\end{table}
%----------------------------------------------
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{平行语料}
\parinterval 神经机器翻译系统的训练需要大量的双语数据，这里我们汇总了一些公开的平行语料，方便读者获取。
\vspace{0.5em}
\begin{itemize}
\item News Commentary Corpus：包括汉语、英语等12个语种，64个语言对的双语数据，爬取自Project Syndicate网站的政治、经济评论。URL：\url{http://www.casmacat.eu/corpus/news-commentary.html}
\vspace{0.5em}
\item CWMT Corpus：中国计算机翻译研讨会社区收集和共享的中英平行语料，涵盖多种领域，例如新闻、电影字幕、小说和政府文档等。URL：\url{http://nlp.nju.edu.cn/cwmt-wmt/}
\vspace{0.5em}
\item Common Crawl corpus：包括捷克语、德语、俄语、法语4种语言到英语的双语数据，爬取自互联网网页。URL：\url{http://www.statmt.org/wmt13/training-parallel-commoncrawl.tgz}
\vspace{0.5em}
\item Europarl Corpus：包括保加利亚语、捷克语等20种欧洲语言到英语的双语数据，来源于欧洲议会记录。URL：\url{http://www.statmt.org/europarl/}
\vspace{0.5em}
\item ParaCrawl Corpus：包括23种欧洲语言到英语的双语语料，数据来源于网络爬取。URL：\url{https://www.paracrawl.eu/index.php}
\vspace{0.5em}
\item United Nations Parallel Corpus：包括阿拉伯语、英语、西班牙语、法语、俄语、汉语6种联合国正式语言，30种语言对的双语数据，来源自联合国公共领域的官方记录和其他会议文件。URL：\url{https://conferences.unite.un.org/UNCorpus/}
\vspace{0.5em}
\item TED Corpus：TED大会演讲在其网站公布了自2007年以来的演讲字幕，以及超过100种语言的翻译版本。WIT收集整理了这些数据，以方便科研工作者使用，同时，会为每年的IWSLT评测比赛提供评测数据集。URL：\url{https://wit3.fbk.eu/}
\vspace{0.5em}
\item OpenSubtile：由P. Lison和J. Tiedemann收集自opensubtiles电影字幕网站，包含62种语言、1782个语种对的平行语料，资源相对比较丰富。URL：\url{http://opus.nlpl.eu/OpenSubtitles2018.php}
\vspace{0.5em}
\item Wikititles Corpus：包括古吉拉特语等14个语种，11个语言对的双语数据，数据来源自维基百科的标题。URL：\url{http://data.statmt.org/wikititles/v1/}
\vspace{0.5em}
\item CzEng:捷克语和英语的平行语料，数据来源于欧洲法律、信息技术和小说领域。URL:\url{ http://ufal.mff.cuni.cz/czeng/czeng17}
\vspace{0.5em}
\item Yandex Corpus：俄语和英语的平行语料，爬取自互联网网页。URL：\url{https://translate.yandex.ru/corpus}
\vspace{0.5em}
\item Tilde MODEL Corpus：欧洲语言的多语言开放数据，包含多个数据集，数据来自于经济、新闻、政府、旅游等门户网站。URL：\url{https://tilde-model.s3-eu-west-1.amazonaws.com/Tilde_MODEL_Corpus.html}
\vspace{0.5em}
\item Setimes Corpus：包括克罗地亚语、阿尔巴尼亚等9种巴尔干语言，72个语言对的双语数据，来源于东南欧时报的新闻报道。URL：\url{http://www.statmt.org/setimes/}
\vspace{0.5em}
\item TVsub：收集自电视剧集字幕的中英文对话语料库，包含超过200万的句对，可用于对话领域和长距离上下文信息的研究。URL：\url{https://github.com/longyuewangdcu/tvsub}
\vspace{0.5em}
\item Recipe Corpus：由Cookpad公司创建的日英食谱语料库，包含10万多的句对。URL：\url{http://lotus.kuee.kyoto-u.ac.jp/WAT/recipe-corpus/}
\end{itemize}
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{相关工具}
%----------------------------------------------------------------------------------------
% NEW SUB-SECTION
%----------------------------------------------------------------------------------------
\subsection{数据预处理工具}
\parinterval 数据处理是搭建神经机器翻译系统的重要步骤，这里我们提供了一些开源工具供读者进行使用。
\vspace{0.5em}
\begin{itemize}
\item Moses：Moses 提供了很多数据预处理的脚本和工具，被机器翻译研究者广泛使用。其中包括符号标准化、分词、大小写转换和长度过滤等。URL：\url{https://github.com/moses-smt/mosesdecoder/tree/master/scripts}
\vspace{0.5em}
\item Jieba：常用的中文分词工具。URL：\url{https://github.com/fxsjy/jieba}
\vspace{0.5em}
\item Subword-nmt：基于BPE算法的子词切分工具。URL：\url{https://github.com/rsennrich/subword-nmt}
\end{itemize}
%----------------------------------------------------------------------------------------
% NEW SUB-SECTION
%----------------------------------------------------------------------------------------
\subsection{评价工具}
\parinterval 机器翻译领域已经有多种自动评价指标，包括BLEU、TER和METEOR等，这里我们提供了一些自动评价指标的工具，方便读者使用。
\vspace{0.5em}
\begin{itemize}
\item Moses：其中包括了通用的BLEU评测脚本。URL：\url{https://github.com/moses-smt/mosesdecoder/tree/master/scripts/generic}
\vspace{0.5em}
\item Tercom：自动评价指标TER的计算工具，只有java版本。URL：\url{http://www.cs.umd.edu/~snover/tercom/}
\vspace{0.5em}
\item Meteor：自动评价指标METEOR的实现。URL：\url{https://www.cs.cmu.edu/~alavie/METEOR/}
\end{itemize}
\end{appendices}
%----------------------------------------------------------------------------------------
% CHAPTER APPENDIX B
%----------------------------------------------------------------------------------------
\begin{appendices}
\chapter{附录B}
\label{appendix-B}
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{IBM模型3训练方法}
\parinterval 模型3的参数估计与模型1和模型2采用相同的方法。这里直接给出辅助函数。
\begin{eqnarray}
h(t,d,n,p, \lambda,\mu, \nu, \zeta) & = & \textrm{P}_{\theta}(\mathbf{s}|\mathbf{t})-\sum_{t}\lambda_{t}\big(\sum_{s}t(s|t)-1\big) \nonumber \\
& & -\sum_{i}\mu_{iml}\big(\sum_{j}d(j|i,m,l)-1\big) \nonumber \\
& & -\sum_{t}\nu_{t}\big(\sum_{\varphi}n(\varphi|t)-1\big)-\zeta(p^0+p^1-1)
\label{eq:1.1}
\end{eqnarray}
\parinterval 由于篇幅所限这里略去了推导步骤直接给出一些用于参数估计的等式。
\begin{eqnarray}
c(s|t,\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times \sum_{j=1}^{m} (\delta(s_j,s) \cdot \delta(t_{a_{j}},t))\big] \label{eq:1.2} \\
c(j|i,m,l;\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times \delta(i,a_j)\big] \label{eq:1.3} \\
c(\varphi|t;\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times \sum_{i=1}^{l}\delta(\varphi,\varphi_{i})\delta(t,t_i)\big]
\label{eq:1.4}
\end{eqnarray}
\begin{eqnarray}
c(0|\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times (m-2\varphi_0) \big] \label{eq:1.5} \\
c(1|\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times \varphi_0 \big] \label{eq:1.6}
\end{eqnarray}
\parinterval 进一步，对于由$K$个样本组成的训练集，有：
\begin{eqnarray}
t(s|t) & = & \lambda_{t}^{-1} \times \sum_{k=1}^{K}c(s|t;\mathbf{s}^{[k]},\mathbf{t}^{[k]}) \label{eq:1.7} \\
d(j|i,m,l) & = & \mu_{iml}^{-1} \times \sum_{k=1}^{K}c(j|i,m,l;\mathbf{s}^{[k]},\mathbf{t}^{[k]}) \label{eq:1.8} \\
n(\varphi|t) & = & \nu_{t}^{-1} \times \sum_{s=1}^{K}c(\varphi |t;\mathbf{s}^{[k]},\mathbf{t}^{[k]}) \label{eq:1.9} \\
p_x & = & \zeta^{-1} \sum_{k=1}^{K}c(x;\mathbf{s}^{[k]},\mathbf{t}^{[k]}) \label{eq:1.10}
\end{eqnarray}
\parinterval 在模型3中，因为产出率的引入，并不能像模型1和模型2那样，在保证正确性的情况下加速参数估计的过程。这就使得每次迭代过程中，都不得不面对大小为$(l+1)^m$的词对齐空间。遍历所有$(l+1)^m$个词对齐所带来的高时间复杂度显然是不能被接受的。因此就要考虑能否仅利用词对齐空间中的部分词对齐对这些参数进行估计。比较简单且直接的方法就是仅利用Viterbi对齐来进行参数估计\footnote{Viterbi词对齐可以被简单的看作搜索到的最好词对齐。}。 遗憾的是，在模型3中并没有方法直接获得Viterbi对齐。这样只能采用一种折中的策略，即仅考虑那些使得$\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t})$达到较高值的词对齐。这里把这部分词对齐组成的集合记为$S$。式\ref{eq:1.2}可以被修改为：
\begin{eqnarray}
c(s|t,\mathbf{s},\mathbf{t}) \approx \sum_{\mathbf{a} \in \mathbf{S}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times \sum_{j=1}^{m}(\delta(s_j,\mathbf{s}) \cdot \delta(t_{a_{j}},\mathbf{t})) \big]
\label{eq:1.11}
\end{eqnarray}
\parinterval 同理可以获得式\ref{eq:1.3}-\ref{eq:1.6}的修改结果。进一步，在IBM模型3中，可以定义$S$如下：
\begin{eqnarray}
S = N(b^{\infty}(V(\mathbf{s}|\mathbf{t};2))) \cup (\mathop{\cup}\limits_{ij} N(b_{i \leftrightarrow j}^{\infty}(V_{i \leftrightarrow j}(\mathbf{s}|\mathbf{t},2))))
\label{eq:1.12}
\end{eqnarray}
\parinterval 为了理解这个公式，先介绍几个概念。
\begin{itemize}
\item $V(\mathbf{s}|\mathbf{t})$表示Viterbi词对齐，$V(\mathbf{s}|\mathbf{t},1)$、$V(\mathbf{s}|\mathbf{t},2)$和$V(\mathbf{s}|\mathbf{t},3)$就分别对应了模型1、2 和3 的Viterbi 词对齐；
\item 把那些满足第$j$个源语言单词对应第$i$个目标语言单词（$a_j=i$）的词对齐构成的集合记为$\mathbf{A}_{i \leftrightarrow j}(\mathbf{s},\mathbf{t})$。通常称这些对齐中$j$和$i$被``钉''在了一起。在$\mathbf{A}_{i \leftrightarrow j}(\mathbf{s},\mathbf{t})$中使$\textrm{P}(\mathbf{a}|\mathbf{s},\mathbf{t})$达到最大的那个词对齐被记为$V_{i \leftrightarrow j}(\mathbf{s},\mathbf{t})$；
\item 如果两个词对齐，通过交换两个词对齐连接就能互相转化，则称它们为邻居。一个词对齐$\mathbf{a}$的所有邻居记为$N(\mathbf{a})$。
\end{itemize}
\vspace{0.5em}
\parinterval 公式\ref{eq:1.12}中，$b^{\infty}(V(\mathbf{s}|\mathbf{t};2))$ 和 $b_{i \leftrightarrow j}^{\infty}(V_{i \leftrightarrow j}(\mathbf{s}|\mathbf{t},2))$ 分别是对 $V(\mathbf{s}|\mathbf{t};3)$ 和 $V_{i \leftrightarrow j}(\mathbf{s}|\mathbf{t},3)$ 的估计。在计算$S$的过程中，需要知道一个对齐$\bf{a}$的邻居$\bf{a}^{'}$的概率，即通过$\textrm{P}_{\theta}(\mathbf{a},\mathbf{s}|\mathbf{t})$计算$\textrm{p}_{\theta}(\mathbf{a}',\mathbf{s}|\mathbf{t})$。在模型3中，如果$\bf{a}$和$\bf{a}'$仅区别于某个源语单词对齐到的目标位置上（$a_j \neq a_{j}'$），那么
\begin{eqnarray}
\textrm{P}_{\theta}(\mathbf{a}',\mathbf{s}|\mathbf{t}) & = & \textrm{P}_{\theta}(\mathbf{a},\mathbf{s}|\mathbf{t}) \cdot \nonumber \\
& & \frac{\varphi_{i'}+1}{\varphi_i} \cdot \frac{n(\varphi_{i'}+1|t_{i'})}{n(\varphi_{i'}|t_{i'})} \cdot \frac{n(\varphi_{i}-1|t_{i})}{n(\varphi_{i}|t_{i})} \cdot \nonumber \\
& & \frac{t(s_j|t_{i'})}{t(s_{j}|t_{i})} \cdot \frac{d(j|i',m,l)}{d(j|i,m,l)}
\label{eq:1.13}
\end{eqnarray}
\parinterval 如果$\bf{a}$和$\bf{a}'$区别于两个位置$j_1$和$j_2$的对齐上，$a_{j_{1}}=a_{j_{2}^{'}}$且$a_{j_{2}}=a_{j_{1}^{'}}$，那么
\begin{eqnarray}
\textrm{P}_{\theta}(\mathbf{a'},\mathbf{s}|\mathbf{t}) = \textrm{P}_{\theta}(\mathbf{a},\mathbf{s}|\mathbf{t}) \cdot \frac{t(s_{j_{2}}|t_{a_{j_{2}}})}{t(s_{j_{1}}|t_{a_{j_{1}}})} \cdot \frac{d(j_{2}|a_{j_{2}},m,l)}{d(j_{1}|a_{j_{1}},m,l)}
\label{eq:1.14}
\end{eqnarray}
\parinterval 相比整个词对齐空间，$S$只是一个非常小的子集，因此运算复杂度可以被大大降低。可以看到，模型3的参数估计过程是建立在模型1和模型2的参数估计结果上的。这不仅是因为模型3要利用模型2的Viterbi对齐，而且还因为模型3参数的初值也要直接利用模型2的参数。从这个角度说，模型1，2，3是有序的且向前依赖的。单独的对模型3的参数进行估计是极其困难的。实际上IBM的模型4和模型5也具有这样的性质，即它们都可以利用前一个模型参数估计的结果作为自身参数的初始值。
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{IBM模型4训练方法}
\parinterval 模型4的参数估计基本与模型3一致。需要修改的是扭曲度的估计公式，对于目标语第$i$个cept.生成的第一单词，可以得到（假设有$K$个训练样本）：
\begin{eqnarray}
d_1(\Delta_j|ca,cb;\mathbf{s},\mathbf{t}) = \mu_{1cacb}^{-1} \times \sum_{k=1}^{K}c_1(\Delta_j|ca,cb;\mathbf{s}^{[k]},\mathbf{t}^{[k]})
\label{eq:1.15}
\end{eqnarray}
其中，
\begin{eqnarray}
c_1(\Delta_j|ca,cb;\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{P}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times s_1(\Delta_j|ca,cb;\mathbf{a},\mathbf{s},\mathbf{t})\big] \label{eq:1.16} \\
s_1(\Delta_j|ca,cb;\rm{a},\mathbf{s},\mathbf{t}) & = & \sum_{i=1}^l \big[\varepsilon(\phi_i) \cdot \delta(\pi_{i1}-\odot _{i},\Delta_j) \cdot \nonumber \\
& & \delta(A(t_{i-1}),ca) \cdot \delta(B(\tau_{i1}),cb) \big] \label{eq:1.17}
\end{eqnarray}
且
\begin{eqnarray}
\varepsilon(x) = \begin{cases}
0 & x \leq 0 \\
1 & x > 0
\end{cases}
\label{eq:1.21}
\end{eqnarray}
对于目标语第$i$个cept.生成的其他单词（非第一个单词），可以得到：
\begin{eqnarray}
d_{>1}(\Delta_j|cb;\mathbf{s},\mathbf{t}) = \mu_{>1cb}^{-1} \times \sum_{k=1}^{K}c_{>1}(\Delta_j|cb;\mathbf{s}^{[k]},\mathbf{t}^{[k]})
\label{eq:1.18}
\end{eqnarray}
其中，
\begin{eqnarray}
c_{>1}(\Delta_j|cb;\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\big[\textrm{p}_{\theta}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times s_{>1}(\Delta_j|cb;\mathbf{a},\mathbf{s},\mathbf{t}) \big] \label{eq:1.19} \\
s_{>1}(\Delta_j|cb;\mathbf{a},\mathbf{s},\mathbf{t}) & = & \sum_{i=1}^l \big[\varepsilon(\phi_i-1)\sum_{k=2}^{\phi_i}\delta(\pi_{[i]k}-\pi_{[i]k-1},\Delta_j) \cdot \nonumber ß\\
& & \delta(B(\tau_{[i]k}),cb) \big] \label{eq:1.20}
\end{eqnarray}
\noindent 这里，$ca$和$cb$分别表示目标语言和源语言的某个词类。模型4需要像模型3一样，通过定义一个词对齐集合$S$，使得每次迭代都在$S$上进行，进而降低运算量。模型4中$S$的定义为：
\begin{eqnarray}
\textrm{S} = N(\tilde{b}^{\infty}(V(\mathbf{s}|\mathbf{t};2))) \cup (\mathop{\cup}\limits_{ij} N(\tilde{b}_{i \leftrightarrow j}^{\infty}(V_{i \leftrightarrow j}(\mathbf{s}|\mathbf{t},2))))
\label{eq:1.22}
\end{eqnarray}
\parinterval 对于一个对齐$\mathbf{a}$，可用模型3对它的邻居进行排名，即按$\textrm{P}_{\theta}(b(\mathbf{a})|\mathbf{s},\mathbf{t};3)$排序，其中$b(\mathbf{a})$表示$\mathbf{a}$的邻居。$\tilde{b}(\mathbf{a})$ 表示这个排名表中满足$\textrm{P}_{\theta}(\mathbf{a}'|\mathbf{s},\mathbf{t};4) > \textrm{P}_{\theta}⁡(\mathbf{a}|\mathbf{s},\mathbf{t};4)$的最高排名的$\mathbf{a}'$。同理可知$\tilde{b}_{i \leftrightarrow j}^{\infty}(\mathbf{a})$ 的意义。这里之所以不用模型3中采用的方法直接利用$b^{\infty}(\mathbf{a})$得到模型4中高概率的对齐，是因为模型4中，要想获得某个对齐$\mathbf{a}$的邻居$\mathbf{a}'$，必须做很大调整，比如：调整$\tau_{[i]1}$和$\odot_{i}$等等。这个过程要比模型3的相应过程复杂得多。因此在模型4中只能借助于模型3的中间步骤来进行参数估计。
\setlength{\belowdisplayskip}{3pt}%调整空白大小
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{IBM模型5训练方法}
\parinterval 模型5的参数估计过程也与模型3的过程基本一致，二者的区别在于扭曲度的估计公式。在模型5中，对于目标语第$i$个cept.生成的第一单词，可以得到（假设有$K$个训练样本）：
\begin{eqnarray}
d_1(\Delta_j|cb;\mathbf{s},\mathbf{t}) = \mu_{1cb}^{-1} \times \sum_{k=1}^{K}c_1(\Delta_j|cb;\mathbf{s}^{[k]},\mathbf{t}^{[k]})
\label{eq:1.23}
\end{eqnarray}
其中，
\begin{eqnarray}
c_1(\Delta_j|cb,v_x,v_y;\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\Big[ \textrm{P}(\mathbf{s},\mathbf{a}|\mathbf{t}) \times s_1(\Delta_j|cb,v_x,v_y;\mathbf{a},\mathbf{s},\mathbf{t}) \Big] \label{eq:1.24} \\
s_1(\Delta_j|cb,v_x,v_y;\mathbf{a},\mathbf{s},\mathbf{t}) & = & \sum_{i=1}^l \Big [ \varepsilon(\phi_i) \cdot \delta(v_{\pi_{i1}},\Delta_j) \cdot \delta(v_{\odot _{i-1}},v_x) \nonumber \\
& & \cdot \delta(v_m-\phi_i+1,v_y) \cdot \delta(v_{\pi_{i1}},v_{\pi_{i1}-1} )\Big] \label{eq:1.25}
\end{eqnarray}
对于目标语第$i$个cept.生成的其他单词（非第一个单词），可以得到：
\begin{eqnarray}
d_{>1}(\Delta_j|cb,v;\mathbf{s},\mathbf{t}) = \mu_{>1cb}^{-1} \times \sum_{k=1}^{K}c_{>1}(\Delta_j|cb,v;\mathbf{s}^{[k]},\mathbf{t}^{[k]})
\label{eq:1.26}
\end{eqnarray}
其中，
\begin{eqnarray}
c_{>1}(\Delta_j|cb,v;\mathbf{s},\mathbf{t}) & = & \sum_{\mathbf{a}}\Big[\textrm{P}(\mathbf{a},\mathbf{s}|\mathbf{t}) \times s_{>1}(\Delta_j|cb,v;\mathbf{a},\mathbf{s},\mathbf{t}) \Big] \label{eq:1.27} \\
s_{>1}(\Delta_j|cb,v;\mathbf{a},\mathbf{s},\mathbf{t}) & = & \sum_{i=1}^l\Big[\varepsilon(\phi_i-1)\sum_{k=2}^{\phi_i} \big[\delta(v_{\pi_{ik}}-v_{\pi_{[i]k}-1},\Delta_j) \nonumber \\
& & \cdot \delta(B(\tau_{[i]k}) ,cb) \cdot \delta(v_m-v_{\pi_{i(k-1)}}-\phi_i+k,v) \nonumber \\
& & \cdot \delta(v_{\pi_{i1}},v_{\pi_{i1}-1}) \big] \Big] \label{eq:1.28}
\end{eqnarray}
\vspace{0.5em}
\parinterval 从式(\ref{eq:1.24})中可以看出因子$\delta(v_{\pi_{i1}},v_{\pi_{i1}-1})$保证了，即使对齐$\mathbf{a}$不合理（一个源语位置对应多个目标语位置）也可以避免在这个不合理的对齐上计算结果。需要注意的是因子$\delta(v_{\pi_{p1}},v_{\pi_{p1-1}})$，确保了$\mathbf{a}$中不合理的部分不产生坏的影响，而$\mathbf{a}$中其他正确的部分仍会参与迭代。
\parinterval 不过上面的参数估计过程与IBM前4个模型的参数估计过程并不完全一样。IBM前4个模型在每次迭代中，可以在给定$\mathbf{s}$、$\mathbf{t}$和一个对齐$\mathbf{a}$的情况下直接计算并更新参数。但是在模型5的参数估计过程中（如公式\ref{eq:1.24}），需要模拟出由$\mathbf{t}$生成$\mathbf{s}$的过程才能得到正确的结果，因为从$\mathbf{t}$、$\mathbf{s}$和$\mathbf{a}$中是不能直接得到 的正确结果的。具体说，就是要从目标语言句子的第一个单词开始到最后一个单词结束，依次生成每个目标语言单词对应的源语言单词，每处理完一个目标语言单词就要暂停，然后才能计算式\ref{eq:1.24}中求和符号里面的内容。这也就是说即使给定了$\mathbf{s}$、$\mathbf{t}$和一个对齐$\mathbf{a}$，也不能直接在它们上进行计算，必须重新模拟$\mathbf{t}$到$\mathbf{s}$的生成过程。
\parinterval 从前面的分析可以看出，虽然模型5比模型4更精确，但是模型5过于复杂以至于给参数估计增加了计算量（对于每组$\mathbf{t}$、$\mathbf{s}$和$\mathbf{a}$都要模拟$\mathbf{t}$生成$\mathbf{s}$的翻译过程）。因此模型5的开发对于系统实现是一个挑战。
\parinterval 在模型5中同样需要定义一个词对齐集合$S$，使得每次迭代都在$S$上进行。可以对$S$进行如下定义
\begin{eqnarray}
\textrm{S} = N(\tilde{\tilde{b}}^{\infty}(V(\mathbf{s}|\mathbf{t};2))) \cup (\mathop{\cup}\limits_{ij} N(\tilde{\tilde{b}}_{i \leftrightarrow j}^{\infty}(V_{i \leftrightarrow j}(\mathbf{s}|\mathbf{t},2))))
\label{eq:1.29}
\end{eqnarray}
\vspace{0.5em}
\parinterval 这里$\tilde{\tilde{b}}(\mathbf{a})$借用了模型4中$\tilde{b}(\mathbf{a})$的概念。不过$\tilde{\tilde{b}}(\mathbf{a})$表示在利用模型3进行排名的列表中满足$\textrm{P}_{\theta}(\mathbf{a}'|\mathbf{s},\mathbf{t};5)$的最高排名的词对齐。
\end{appendices}
ChapterPreface/Figures/figure-preface.tex 0 → 100644
% !Mode:: "TeX:UTF-8"
% !TEX encoding = UTF-8 Unicode
\begin{tikzpicture}
\tikzstyle{secnode} =[font=\scriptsize,minimum height=4.0em,minimum width=22em,draw,thick,fill=white,drop shadow]
\tikzstyle{conceptnode} =[font=\scriptsize,minimum height=1.5em,minimum width=5em]
\tikzstyle{conceptnodesmall} =[font=\scriptsize,minimum height=1.0em,minimum width=4.4em]
% section 1
\node [secnode,anchor=south west,minimum width=22.5em,red,fill=white] (sec1) at (0,0) {};
\node [anchor=north] (sec1label) at ([yshift=-0.2em]sec1.north) {\small{机器翻译简介}};
\node [anchor=north west,draw=red,thick,fill=white,rounded corners] (sec1title) at ([xshift=-0.3em,yshift=0.3em]sec1.north west) {{\footnotesize\bfnew{\color{red} 第一章}}};
\node [conceptnode,anchor=south west,fill=red!15,thin] (sec1box1) at ([xshift=0.5em,yshift=0.5em]sec1.south west) {\footnotesize{发展历史}};
\node [conceptnode,anchor=west,fill=red!15,thin] (sec1box2) at ([xshift=0.5em]sec1box1.east) {\footnotesize{评价方法}};
\node [conceptnode,anchor=west,fill=red!15,thin] (sec1box3) at ([xshift=0.5em]sec1box2.east) {\footnotesize{应用情况}};
\node [conceptnode,anchor=west,fill=red!15,thin] (sec1box4) at ([xshift=0.5em]sec1box3.east) {\footnotesize{系统\&数据}};
% section 2
\node [secnode,anchor=south,blue,fill=white] (sec2) at ([xshift=-6.5em,yshift=3em]sec1.north) {};
\node [anchor=north] (sec2label) at (sec2.north) {\small{词法、语法及统计建模基础}};
\node [anchor=north west,draw=blue,thick,fill=white,rounded corners] (sec2title) at ([xshift=-0.3em,yshift=0.3em]sec2.north west) {{\footnotesize\bfnew{\color{blue} 第二章}}};
\node [conceptnode,anchor=south west,fill=ublue!15,thin,minimum width=4em,align=left] (sec2box1) at ([xshift=0.5em,yshift=0.4em]sec2.south west) {\tiny{概率论与}\\\tiny{统计建模基础}};
\node [anchor=west,draw,dotted,thick,minimum height=2em,minimum width=16.6em,align=left] (sec2box2) at ([xshift=0.3em]sec2box1.east) {};
\node [conceptnodesmall,minimum width=5em,anchor=south west,fill=blue!15,thin] (sec2box3) at ([xshift=0.4em,yshift=0.3em]sec2box2.south west) {\scriptsize{中文分词}};
\node [conceptnodesmall,minimum width=5em,anchor=west,fill=blue!15,thin] (sec2box4) at ([xshift=0.4em]sec2box3.east) {\scriptsize{$n$元语法模型}};
\node [conceptnodesmall,minimum width=5em,anchor=west,fill=blue!15,thin] (sec2box5) at ([xshift=0.4em]sec2box4.east) {\scriptsize{句法分析}};
\draw [->,very thick] ([xshift=-1em,yshift=0.2em]sec1.north) .. controls +(north:2.5em) and +(south:2.5em) .. ([xshift=-3em,yshift=-0.2em]sec2.south);
% section 5
\node [secnode,anchor=south,orange,fill=white] (sec5) at ([xshift=7em,yshift=10em]sec1.north) {};
\node [anchor=north] (sec5label) at (sec5.north) {\small{人工神经网络和神经语言建模}};
\node [anchor=north west,draw=orange,thick,fill=white,rounded corners] (sec5title) at ([xshift=-0.3em,yshift=0.3em]sec5.north west) {{\footnotesize\bfnew{\color{orange} 第五章}}};
\node [conceptnode,minimum width=4em,anchor=south west,fill=orange!15,thin,align=left] (sec5box1) at ([xshift=0.5em,yshift=0.4em]sec5.south west) {\tiny{线性代数基础}\\\tiny{与感知机}};
\node [conceptnode,minimum width=4em,anchor=west,fill=orange!15,thin,align=left] (sec5box2) at ([xshift=0.2em]sec5box1.east) {\tiny{多层神经网络}\\\tiny{与实现方法}};
\node [conceptnode,minimum width=4em,anchor=west,fill=orange!15,thin,align=left] (sec5box3) at ([xshift=0.2em]sec5box2.east) {\tiny{模型训练}\\\tiny{(反向传播)}};
\node [conceptnode,minimum width=4em,anchor=west,fill=purple!15,thin,align=left] (sec5box4) at ([xshift=0.2em]sec5box3.east) {\tiny{神经语言模型}\\\tiny{(FNN等)}};
\node [conceptnode,minimum width=4em,anchor=west,fill=purple!15,thin,align=left] (sec5box5) at ([xshift=0.2em]sec5box4.east) {\tiny{表示学习与}\\\tiny{预训练模型}};
\node [draw,dotted,thick,inner sep=1pt] [fit = (sec5box4) (sec5box5)] (pretrainbox) {};
\draw [->,very thick] ([yshift=-9.8em]sec5.south) -- ([yshift=-0.2em]sec5.south);
\draw [->,thick,dotted] ([xshift=0.2em,yshift=1em]sec2.east) .. controls +(east:3em) and +(south:4em) .. ([xshift=3em,yshift=-0.0em]pretrainbox.south);
% section 3
\node [secnode,anchor=south,ugreen,fill=white] (sec3) at ([yshift=10em]sec2.north) {};
\node [anchor=north] (sec3label) at (sec3.north) {\small{基于词的机器翻译模型}};
\node [anchor=north west,draw=ugreen,thick,fill=white,rounded corners] (sec3title) at ([xshift=-0.3em,yshift=0.3em]sec3.north west) {{\footnotesize\bfnew{\color{ugreen} 第三章}}};
\node [conceptnode,minimum width=4em,anchor=south west,fill=ublue!15,thin,align=left] (sec3box1) at ([xshift=0.5em,yshift=0.4em]sec3.south west) {\tiny{机器翻译的统计}\\\tiny{描述（实例）}};
\node [conceptnode,minimum width=4em,anchor=west,fill=green!20,thin,align=left] (sec3box2) at ([xshift=0.2em]sec3box1.east) {\tiny{噪声信道模型}\\\tiny{与生成式模型}};
\node [conceptnode,minimum width=4em,anchor=west,fill=green!20,thin,align=left] (sec3box3) at ([xshift=0.2em]sec3box2.east) {\tiny{IBM模型、隐}\\\tiny{马尔可夫模型}};
\node [conceptnode,minimum width=4em,anchor=west,fill=green!20,thin,align=left] (sec3box4) at ([xshift=0.2em]sec3box3.east) {\tiny{\hspace{0.9em}参数学习}\\\tiny{=优化}};
\node [conceptnode,minimum width=3.8em,anchor=west,fill=green!20,thin,align=left,minimum height=2em,inner sep=2pt] (sec3box5) at ([xshift=0.2em]sec3box4.east) {\scriptsize{EM算法}};
\draw [->,very thick] ([yshift=0.2em,xshift=-3em]sec2.north) -- ([yshift=-0.2em,xshift=-3em]sec3.south);
% section 4
\node [secnode,anchor=south,ugreen,fill=white] (sec4) at ([yshift=3em]sec3.north) {};
\node [anchor=north] (sec4label) at (sec4.north) {\small{基于短语和句法的机器翻译模型}};
\node [anchor=north west,draw=ugreen,thick,fill=white,rounded corners] (sec4title) at ([xshift=-0.3em,yshift=0.3em]sec4.north west) {{\footnotesize\bfnew{\color{ugreen} 第四章}}};
\node [conceptnode,minimum width=4em,anchor=south west,fill=ublue!15,thin,align=left] (sec4box1) at ([xshift=0.5em,yshift=0.4em]sec4.south west) {\tiny{判别式模型与}\\\tiny{最小错误率训练}};
\node [conceptnode,minimum width=4em,anchor=west,fill=green!20,thin,align=left] (sec4box2) at ([xshift=0.2em]sec4box1.east) {\tiny{基于翻译推导}\\\tiny{的建模}};
\node [conceptnode,minimum width=4em,anchor=west,fill=green!20,thin,align=left] (sec4box3) at ([xshift=0.2em]sec4box2.east) {\tiny{短语及句法}\\\tiny{翻译规则抽取}};
\node [conceptnode,minimum width=4em,anchor=west,fill=green!20,thin,align=left,minimum height=2em,inner sep=2pt] (sec4box4) at ([xshift=0.2em]sec4box3.east) {\scriptsize{调序模型}};
\node [conceptnode,minimum width=3.8em,anchor=west,fill=green!20,thin,align=left,minimum height=2em,inner sep=2pt] (sec4box5) at ([xshift=0.2em]sec4box4.east) {\scriptsize{解码}};
\draw [->,very thick] ([yshift=0.2em,xshift=-3em]sec3.north) -- ([yshift=-0.2em,xshift=-3em]sec4.south);
% section 6
\node [secnode,anchor=south,purple,fill=white] (sec6) at ([yshift=19em]sec5.north) {};
\node [anchor=north] (sec6label) at (sec6.north) {\small{神经机器翻译模型}};
\node [anchor=north west,draw=purple,thick,fill=white,rounded corners] (sec6title) at ([xshift=-0.3em,yshift=0.3em]sec6.north west) {{\footnotesize\bfnew{\color{purple} 第六章}}};
\node [conceptnode,minimum width=4em,anchor=south west,fill=ublue!15,thin,align=left] (sec6box1) at ([xshift=0.5em,yshift=0.4em]sec6.south west) {\tiny{编码器-解码器}\\\tiny{框架}};
\node [conceptnode,minimum width=4em,anchor=west,fill=ublue!15,thin,align=left,minimum height=2em] (sec6box2) at ([xshift=0.2em]sec6box1.east) {\scriptsize{注意力机制}};
\node [conceptnode,minimum width=7.5em,anchor=west,fill=purple!15,thin,align=left] (sec6box3) at ([xshift=0.2em]sec6box2.east) {\tiny{基于RNN和Transformer}\\\tiny{的神经机器翻译建模}};
\node [conceptnode,minimum width=4em,anchor=west,fill=purple!15,thin,align=left,minimum height=2em] (sec6box4) at ([xshift=0.2em]sec6box3.east) {\scriptsize{训练与推断}};
\draw [->,very thick] ([yshift=0.2em]sec5.north) -- ([yshift=-0.2em]sec6.south);
\draw [->,very thick,dotted] ([yshift=0.2em,xshift=-2em]sec4.north) .. controls +(north:5.0em) and +(west:4em) .. ([xshift=-0.2em]sec6.west);
\draw [->,thick,dotted] ([xshift=3em,yshift=0.2em]pretrainbox.north) .. controls +(north:15em) and +(south:15em) .. ([xshift=0em,yshift=-0.0em]sec6box3.south);
% section 7
\node [secnode,anchor=south,purple,fill=white,minimum height=6.3em] (sec7) at ([yshift=3em]sec6.north) {};
\node [anchor=north] (sec7label) at (sec7.north) {\small{神经机器翻译实战}};
\node [anchor=north west,draw=purple,thick,fill=white,rounded corners] (sec7title) at ([xshift=-0.3em,yshift=0.3em]sec7.north west) {{\footnotesize\bfnew{\color{purple} 第七章}}};
\node [conceptnode,minimum width=4em,anchor=south west,fill=ublue!15,thin,align=left,minimum height=4.2em] (sec7box1) at ([xshift=0.5em,yshift=0.4em]sec7.south west) {\tiny{数据处理、}\\\tiny{子词切分}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box2) at ([xshift=0.5em,yshift=-0.2em]sec7box1.north east) {\tiny{正则化}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box3) at ([yshift=-0.1em]sec7box2.south west) {\tiny{增大模型容量}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box4) at ([yshift=-0.1em]sec7box3.south west) {\tiny{大批量训练}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box5) at ([xshift=0.6em]sec7box2.north east) {\tiny{推断优化}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box6) at ([yshift=-0.1em]sec7box5.south west) {\tiny{译文长度控制}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box7) at ([yshift=-0.1em]sec7box6.south west) {\tiny{多模型集成}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box8) at ([xshift=0.6em]sec7box5.north east) {\tiny{深层模型}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box9) at ([yshift=-0.1em]sec7box8.south west) {\tiny{知识精炼}};
\node [anchor=north west,minimum width=5em,anchor=north west,fill=purple!15] (sec7box10) at ([yshift=-0.1em]sec7box9.south west) {\tiny{单语数据使用}};
\node [draw,dotted,thick,inner sep=1pt] [fit = (sec7box2) (sec7box3) (sec7box4)] (trainbox) {};
\node [draw,dotted,thick,inner sep=1pt] [fit = (sec7box5) (sec7box6) (sec7box7)] (inferencebox) {};
\node [draw,dotted,thick,inner sep=1pt] [fit = (sec7box8) (sec7box9) (sec7box10)] (advancedbox) {};
\draw [->,very thick] ([yshift=0.2em]sec6.north) -- ([yshift=-0.2em]sec7.south);
\draw [->,very thick,dotted] ([yshift=0.2em,xshift=-3em]sec4.north) .. controls +(north:7.0em) and +(west:6em) .. ([xshift=-0.2em]sec7.west);
%caption
\node [anchor=north] (caption) at ([xshift=0.4em,yshift=-1em]sec1.south) {\footnotesize{本书各章节及核心概念关系图}};
\end{tikzpicture}
ChapterPreface/chapterpreface.tex 0 → 100644
% !Mode:: "TeX:UTF-8"
% !TEX encoding = UTF-8 Unicode
%----------------------------------------------------------------------------------------
% 机器翻译：统计建模与深度学习方法
% Machine Translation: Statistical Modeling and Deep Learning Methods
%
% Copyright 2020
% 肖桐(xiaotong@mail.neu.edu.cn) 朱靖波 (zhujingbo@mail.neu.edu.cn)
%----------------------------------------------------------------------------------------
\renewcommand\figurename{图}
%----------------------------------------------------------------------------------------
% PREFACE
%----------------------------------------------------------------------------------------
{\color{white} 空}
\vspace{0.5em}
\begin{center}
{\Huge \bfnew{导\ \ \ \ 读}}
\end{center}
\vspace{2em}
\begin{spacing}{1.18}
让计算机进行自然语言的翻译是人类长期的梦想，也是人工智能的终极目标之一。自上世纪九十年代起，机器翻译迈入了基于统计建模的时代，发展到今天，深度学习等机器学习方法已经在机器翻译中得到了大量的应用，取得了令人瞩目的进步。
在这个时代背景下，对机器翻译的模型、方法和实现技术进行深入了解是自然语言处理领域研究者和实践者所渴望的。本书全面回顾了近三十年内机器翻译的技术发展历程，并围绕统计建模和深度学习两个主题对机器翻译的技术方法进行了全面介绍。在写作中，笔者力求用朴实的语言和简洁的实例阐述机器翻译的基本模型和方法，同时对相关的技术前沿进行讨论。本书可以供计算机相关专业高年级本科生及研究生学习之用，也可以作为自然语言处理，特别是机器翻译领域相关研究人员的参考资料。
本书共分为七个章节，章节的顺序参考了机器翻译技术发展的时间脉络，同时兼顾了机器翻译知识体系的内在逻辑。各章节的主要内容包括：
\begin{itemize}
\vspace{0.5em}
\item 第一章：机器翻译简介
\vspace{0.5em}
\item 第二章：词法、语法及统计建模基础
\vspace{0.5em}
\item 第三章：基于词的机器翻译模型
\vspace{0.5em}
\item 第四章：基于短语和句法的机器翻译模型
\vspace{0.5em}
\item 第五章：人工神经网络和神经语言建模
\vspace{0.5em}
\item 第六章：神经机器翻译模型
\vspace{0.5em}
\item 第七章：神经机器翻译实战 \ \dash \ 参加一次比赛
\vspace{0.5em}
\end{itemize}
其中，第一章是对机器翻译的整体介绍。第二章和第五章是对统计建模和深度学习方法的介绍，分别建立了两个机器翻译范式的基础知识体系 \ \dash \ 统计机器翻译和神经机器翻译。统计机器翻译部分（第三、四章）涉及早期的基于单词的翻译模型，以及本世纪初流行的基于短语和句法的翻译模型。神经机器翻译（第六、七章）代表了当今机器翻译的前沿，内容主要涉及了基于端到端表示学习的机器翻译建模方法。特别地，第七章对一些最新的神经机器翻译方法进行了讨论，为相关科学问题的研究和实用系统的开发提供了可落地的思路。下图展示了本书各个章节及核心概念之间的关系。
{\red 用最简单的方式阐述机器翻译的基本思想}是笔者所期望达到的目标。但是，书中不可避免会使用一些形式化定义和算法的抽象描述，因此，笔者尽所能通过图例进行解释（本书共320张插图）。不过，本书所包含的内容较为广泛，难免会有疏漏，望读者海涵，并指出不当之处。
%-------------------------------------------
\begin{figure}[htp]
\centering
\centering
\input{./ChapterPreface/Figures/figure-preface}
\end{figure}
%-------------------------------------------
\end{spacing}
Figures/chapter_head_1.pdf 0 → 100644
No preview for this file type
Figures/chapter_head_2.pdf 0 → 100644
This source diff could not be displayed because it is too large. You can view the blob instead.
bibliography.bib 0 → 100644
This source diff could not be displayed because it is too large. You can view the blob instead.
mt-book-xelatex.tex 0 → 100644
% !Mode:: "TeX:UTF-8"
% !TEX encoding = UTF-8 Unicode
%----------------------------------------------------------------------------------------
% 机器翻译：统计建模与深度学习方法
% Machine Translation: Statistical Modeling and Deep Learning Methods
%
% Copyright 2020
% 肖桐(xiaotong@mail.neu.edu.cn) 朱靖波 (zhujingbo@mail.neu.edu.cn)
%----------------------------------------------------------------------------------------
%----------------------------------------------------------------------------------------
% BASIC CONFIGURATIONS
%----------------------------------------------------------------------------------------
\documentclass[11pt]{book} % font and book template
\input{structure.tex} % self-defined template
\usepackage{hyperref}
%\hypersetup{pdftitle={Title},pdfauthor={Author}} % Uncomment and fill out to include PDF metadata for the author and title of the book
\usepackage {xeCJK}
\usepackage{ctex}
\setCJKmainfont{SimSun}
\setCJKmonofont{SimSun}
\setmainfont{Times New Roman}
%----------------------------------------------------------------------------------------
% CHINESE FONTS AND MATH FONTS
%----------------------------------------------------------------------------------------
{\newcommand{\mycfont}{song}}
{\newcommand{\mycfont}{gbsn}}
% math fount = Computer Modern Roman
\AtBeginDocument{
\SetSymbolFont{operators}{normal}{OT1}{cmr} {m}{n}
\SetSymbolFont{letters}{normal}{OML}{cmm} {m}{it}
\SetSymbolFont{symbols}{normal}{OMS}{cmsy}{m}{n}
\SetSymbolFont{largesymbols}{normal}{OMX}{cmex}{m}{n}
\SetSymbolFont{operators}{bold}{OT1}{cmr} {bx}{n}
\SetSymbolFont{letters}{bold}{OML}{cmm} {b}{it}
\SetSymbolFont{symbols}{bold}{OMS}{cmsy}{b}{n}
\SetSymbolFont{largesymbols}{bold}{OMX}{cmex}{m}{n}
\SetMathAlphabet{\mathbf}{normal}{OT1}{cmr}{bx}{n}
\SetMathAlphabet{\mathsf}{normal}{OT1}{cmss}{m}{n}
\SetMathAlphabet{\mathit}{normal}{OT1}{cmr}{m}{it}
\SetMathAlphabet{\mathtt}{normal}{OT1}{cmtt}{m}{n}
\SetMathAlphabet{\mathbf}{bold}{OT1}{cmr}{bx}{n}
\SetMathAlphabet{\mathsf}{bold}{OT1}{cmss}{bx}{n}
\SetMathAlphabet{\mathit}{bold}{OT1}{cmr}{bx}{it}
\SetMathAlphabet{\mathtt}{bold}{OT1}{cmtt}{m}{n}
}
\renewcommand{\baselinestretch}{1.2} % spacing
%----------------------------------------------------------------------------------------
% MAIN BODY OF THE BOOK
%----------------------------------------------------------------------------------------
\begin{document}
%\begin{CJK}{UTF8}{\mycfont}%原来的CJK
%----------------------------------------------------------------------------------------
% TITLE PAGE
%----------------------------------------------------------------------------------------
\begingroup
\thispagestyle{empty}
\begin{tikzpicture}[remember picture,overlay]
\node[inner sep=0pt] (background) at (current page.center) {\includegraphics[width=\paperwidth,height=\paperheight]{fig-cover.jpg}};
\end{tikzpicture}
\vfill
\endgroup·
%----------------------------------------------------------------------------------------
% COPYRIGHT PAGE
%----------------------------------------------------------------------------------------
\newpage
~\vfill
\thispagestyle{empty}
\noindent Copyright \copyright\ 2020 肖桐\ \ 朱靖波\\
\noindent \textsc{东北大学自然语言处理实验室\ $\cdot$\ 小牛翻译}\\
\noindent 顾问：姚天顺\ \ 王宝库\\
\noindent \textsc{\url{https://opensource.niutrans.com/mtbook/index.html}}\\
\noindent \textsc{\url{https://github.com/NiuTrans/MTBook}}\\
\noindent {\red{Licensed under the Creative Commons Attribution-NonCommercial 4.0 Unported License (the ``License''). You may not use this file except in compliance with the License. You may obtain a copy of the License at \url{http://creativecommons.org/licenses/by-nc/4.0}. Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an \textsc{``as is'' basis, without warranties or conditions of any kind}, either express or implied. See the License for the specific language governing permissions and limitations under the License.}}\\
\noindent \textit{\today}
%----------------------------------------------------------------------------------------
% ACKNOWLEDGE PAGE
%----------------------------------------------------------------------------------------
\newpage
~\vfill
\thispagestyle{empty}
{\large
\noindent {\color{red} 在此感谢为本书做出贡献的小牛团队（部分）成员} \\
\noindent 曹润柘、曾信、孟霞、单韦乔、姜雨帆、王子扬、刘辉、许诺、李北、刘继强、张哲旸、周书含、周涛、李炎洋、林野、陈贺轩、刘晓倩、牛蕊、田丰宁、杜权、李垠桥、许晨、张裕浩、胡驰、冯凯、王泽洋、刘腾博、刘兴宇、徐萍、赵闯、高博、张春良、王会珍、张俐、杨木润、宁义明、李洋、秦浩、胡明涵 \\
}
%----------------------------------------------------------------------------------------
% PREFACE PAGES
%----------------------------------------------------------------------------------------
\newpage
\include{ChapterPreface/ChapterPreface}
%----------------------------------------------------------------------------------------
% TABLE OF CONTENTS
%----------------------------------------------------------------------------------------
%\usechapterimagefalse % If you don't want to include a chapter image, use this to toggle images off - it can be enabled later with \usechapterimagetrue
\chapterimage{fig-NEU-1.jpg} % Image of the content page
\pagestyle{empty} % Disable headers and footers for the following pages
\tableofcontents % Show contents
\cleardoublepage % Place the first page of each chapter on odd pages
\pagestyle{fancy} % Enable headers and footers
%----------------------------------------------------------------------------------------
% CHAPTERS
%----------------------------------------------------------------------------------------
\include{Chapter1/chapter1}
\include{Chapter2/chapter2}
\include{Chapter3/chapter3}
\include{Chapter4/chapter4}
\include{Chapter5/chapter5}
\include{Chapter6/chapter6}
\include{Chapter7/chapter7}
\include{ChapterAppend/chapterappend}
%----------------------------------------------------------------------------------------
% BIBLIOGRAPHY
%----------------------------------------------------------------------------------------
\chapterimage{fig-NEU-10.jpg} % Image of the header
\cleardoublepage % Make sure the index starts on an odd (right side) page
\printbibliography
%----------------------------------------------------------------------------------------
% INDEX
%----------------------------------------------------------------------------------------
\chapterimage{fig-NEU-10.jpg} % Image of the header
\cleardoublepage % Make sure the index starts on an odd (right side) page
%\phantomsection
%\setlength{\columnsep}{0.75cm} % Space between the 2 columns of the index
%\addcontentsline{toc}{chapter}{\textcolor{ocre}{Index}} % Add an Index heading to the table of contents
\printindex % Show index
%-------------------------
%\end{CJK}
\end{document}
structure.tex 0 → 100644
% !Mode:: "TeX:UTF-8"
% !TEX encoding = UTF-8 Unicode
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% This file was modified on top of
% The Legrand Orange Book
% Structural Definitions File
%
% Original author:
% Mathias Legrand (legrand.mathias@gmail.com) with modifications by:
% Vel (vel@latextemplates.com)
%
% Current Version is maintained by
% Tong Xiao (xiaotong@mail.neu.edu.cn)
% Runzhe Cao (854581319@qq.com)
%
% License of This File:
% CC BY-NC-SA 4.0 (http://creativecommons.org/licenses/by-nc-sa/4.0/)
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%----------------------------------------------------------------------------------------
% VARIOUS REQUIRED PACKAGES AND CONFIGURATIONS
%----------------------------------------------------------------------------------------
\usepackage{graphicx} % Required for including pictures
\graphicspath{{Figures/}} % Specifies the directory where pictures are stored
\usepackage{lipsum} % Inserts dummy text
\usepackage{tikz} % Required for drawing custom shapes
\usepackage[english]{babel} % English language/hyphenation
\usepackage{enumitem} % Customize lists
\setlist{nolistsep} % Reduce spacing between bullet points and numbered lists
\usepackage{booktabs} % Required for nicer horizontal rules in tables
\usepackage{xcolor} % Required for specifying colors by name
\definecolor{ocre}{RGB}{243,102,25} % Define the orange color used for highlighting throughout the book
%----------------------------------------------------------------------------------------
% MARGINS
%----------------------------------------------------------------------------------------
\usepackage{geometry} % Required for adjusting page dimensions and margins
\geometry{
paper=b5paper, % Paper size, change to letterpaper for US letter size
%papersize={185mm,260mm}, % specify paper size by (width,height)
top=2cm, % Top margin
bottom=1.5cm, % Bottom margin原来1.5cm
left=1.8cm, % Left margin
right=1.8cm, % Right margin
headheight=10pt, % Header height
footskip=1.4cm, % Space from the bottom margin to the baseline of the footer
headsep=10pt, % Space from the top margin to the baseline of the header
%showframe, % Uncomment to show how the type block is set on the page
}
%----------------------------------------------------------------------------------------
% FONTS
%----------------------------------------------------------------------------------------
\usepackage{avant} % Use the Avantgarde font for headings
%\usepackage{times} % Use the Times font for headings
\usepackage{mathptmx} % Use the Adobe Times Roman as the default text font together with math symbols from the Sym­bol, Chancery and Com­puter Modern fonts
\usepackage{microtype} % Slightly tweak font spacing for aesthetics
\usepackage[utf8]{inputenc} % Required for including letters with accents
\usepackage[T1]{fontenc} % Use 8-bit encoding that has 256 glyphs
%----------------------------------------------------------------------------------------
% BIBLIOGRAPHY AND INDEX
%----------------------------------------------------------------------------------------
\usepackage[style=numeric,citestyle=numeric,sorting=anyt,sortcites=true,maxbibnames=40,minbibnames=30,autopunct=true,babel=hyphen,hyperref=true,abbreviate=false,backref=true,backend=biber]{biblatex}
%maxbibnames 设置参考文献最多显示作者数目
%minbibnames 如果作者数目超过maxbibnames，则只显示minbibnames个作者
\addbibresource{bibliography.bib} % BibTeX bibliography file
\defbibheading{bibempty}{}
\usepackage{calc} % For simpler calculation - used for spacing the index letter headings correctly
\usepackage{makeidx} % Required to make an index
\makeindex % Tells LaTeX to create the files required for indexing
\newcommand{\upcite}[1]{\textsuperscript{\textsuperscript{\cite{#1}}}}%参考文献上标
%----------------------------------------------------------------------------------------
% MAIN TABLE OF CONTENTS
%----------------------------------------------------------------------------------------
\usepackage{titletoc} % Required for manipulating the table of contents
\contentsmargin{0cm} % Removes the default margin
% Part text styling (this is mostly taken care of in the PART HEADINGS section of this file)
\titlecontents{part}
[0cm] % Left indentation
{\addvspace{20pt}\bfseries} % Spacing and font options for parts
{}
{}
{}
% Chapter text styling
\titlecontents{chapter}
[1.25cm] % Left indentation
{\addvspace{12pt}\large\sffamily\bfseries} % Spacing and font options for chapters
{\color{ocre!60}\contentslabel[\Large\thecontentslabel]{1.25cm}\color{ocre}} % Formatting of numbered sections of this type
{\color{ocre}} % Formatting of numberless sections of this type
{\color{ocre!60}\normalsize\;\titlerule*[.5pc]{.}\;\thecontentspage} % Formatting of the filler to the right of the heading and the page number
% Section text styling
\titlecontents{section}
[1.25cm] % Left indentation
{\addvspace{3pt}\sffamily\bfseries} % Spacing and font options for sections
{\contentslabel[\thecontentslabel]{1.25cm}} % Formatting of numbered sections of this type
{} % Formatting of numberless sections of this type
{\titlerule*[.5pc]{.}\;\thecontentspage}%
%{\hfill\color{black}\thecontentspage} % Formatting of the filler to the right of the heading and the page number
% Subsection text styling
\titlecontents{subsection}
[1.25cm] % Left indentation
{\addvspace{1pt}\sffamily\small} % Spacing and font options for subsections
{\contentslabel[\thecontentslabel]{1.25cm}} % Formatting of numbered sections of this type
{} % Formatting of numberless sections of this type
{\ \titlerule*[.5pc]{.}\;\thecontentspage} % Formatting of the filler to the right of the heading and the page number
% Figure text styling
\titlecontents{figure}
[1.25cm] % Left indentation
{\addvspace{1pt}\sffamily\small} % Spacing and font options for figures
{\thecontentslabel\hspace*{1em}} % Formatting of numbered sections of this type
{} % Formatting of numberless sections of this type
{\ \titlerule*[.5pc]{.}\;\thecontentspage} % Formatting of the filler to the right of the heading and the page number
% Table text styling
\titlecontents{table}
[1.25cm] % Left indentation
{\addvspace{1pt}\sffamily\small} % Spacing and font options for tables
{\thecontentslabel\hspace*{1em}} % Formatting of numbered sections of this type
{} % Formatting of numberless sections of this type
{\ \titlerule*[.5pc]{.}\;\thecontentspage} % Formatting of the filler to the right of the heading and the page number
%----------------------------------------------------------------------------------------
% MINI TABLE OF CONTENTS IN PART HEADS
%----------------------------------------------------------------------------------------
% Chapter text styling
\titlecontents{lchapter}
[0em] % Left indentation
{\addvspace{15pt}\large\sffamily\bfseries} % Spacing and font options for chapters
{\color{ocre}\contentslabel[\Large\thecontentslabel]{1.25cm}\color{ocre}} % Chapter number
{}
{\color{ocre}\normalsize\sffamily\bfseries\;\titlerule*[.5pc]{.}\;\thecontentspage} % Page number
% Section text styling
\titlecontents{lsection}
[0em] % Left indentation
{\sffamily\small} % Spacing and font options for sections
{\contentslabel[\thecontentslabel]{1.25cm}} % Section number
{}
{}
% Subsection text styling (note these aren't shown by default, display them by searchings this file for tocdepth and reading the commented text)
\titlecontents{lsubsection}
[.5em] % Left indentation
{\sffamily\footnotesize} % Spacing and font options for subsections
{\contentslabel[\thecontentslabel]{1.25cm}}
{}
{}
%----------------------------------------------------------------------------------------
% HEADERS AND FOOTERS
%----------------------------------------------------------------------------------------
\usepackage{fancyhdr} % Required for header and footer configuration
\pagestyle{fancy} % Enable the custom headers and footers
\renewcommand{\chaptermark}[1]{\markboth{\sffamily\normalsize\bfseries\chaptername\ \thechapter.\ #1 \quad 肖桐\ 朱靖波}{}} % Styling for the current chapter in the header
\renewcommand{\sectionmark}[1]{\markright{\sffamily\normalsize\thesection\hspace{5pt}#1}{}} % Styling for the current section in the header
\fancyhf{} % Clear default headers and footers
\fancyhead[LE,RO]{\sffamily\normalsize\thepage} % Styling for the page number in the header
\fancyhead[LO]{\rightmark} % Print the nearest section name on the left side of odd pages
\fancyhead[RE]{\leftmark} % Print the current chapter name on the right side of even pages
%\fancyfoot[RE]{\tiny{肖桐\ 朱靖波}}
%O-odd page,E-even page,R-right area,L-left area,C-center area
%\fancyfoot[C]{\thepage} % Uncomment to include a footer底部中间页码
\renewcommand{\headrulewidth}{0.5pt} % Thickness of the rule under the header
\fancypagestyle{plain}{% Style for when a plain pagestyle is specified
\fancyhead{}\renewcommand{\headrulewidth}{0pt}%
}
% Removes the header from odd empty pages at the end of chapters
\makeatletter
\renewcommand{\cleardoublepage}{
\clearpage\ifodd\c@page\else
\hbox{}
\vspace*{\fill}
\thispagestyle{empty}
\newpage
\fi}
%----------------------------------------------------------------------------------------
% THEOREM STYLES
%----------------------------------------------------------------------------------------
\usepackage{amsmath,amsfonts,amssymb,amsthm} % For math equations, theorems, symbols, etc
\newcommand{\intoo}[2]{\mathopen{]}#1\,;#2\mathclose{[}}
\newcommand{\ud}{\mathop{\mathrm{{}d}}\mathopen{}}
\newcommand{\intff}[2]{\mathopen{[}#1\,;#2\mathclose{]}}
\renewcommand{\qedsymbol}{$\blacksquare$}
\newtheorem{notation}{Notation}[chapter]
% Boxed/framed environments
\newtheoremstyle{ocrenumbox}% Theorem style name
{0pt}% Space above
{0pt}% Space below
{\normalfont}% Body font
{}% Indent amount
{\small\bf\sffamily\color{ocre}}% Theorem head font
{\;}% Punctuation after theorem head
{0.25em}% Space after theorem head
{\small\sffamily\color{ocre}\thmname{#1}\nobreakspace\thmnumber{\@ifnotempty{#1}{}\@upn{#2}}% Theorem text (e.g. Theorem 2.1)
\thmnote{\nobreakspace\the\thm@notefont\sffamily\bfseries\color{black}---\nobreakspace#3.}} % Optional theorem note
\newtheoremstyle{blacknumex}% Theorem style name
{5pt}% Space above
{5pt}% Space below
{\normalfont}% Body font
{} % Indent amount
{\small\bf\sffamily}% Theorem head font
{\;}% Punctuation after theorem head
{0.25em}% Space after theorem head
{\small\sffamily{\tiny\ensuremath{\blacksquare}}\nobreakspace\thmname{#1}\nobreakspace\thmnumber{\@ifnotempty{#1}{}\@upn{#2}}% Theorem text (e.g. Theorem 2.1)
\thmnote{\nobreakspace\the\thm@notefont\sffamily\bfseries---\nobreakspace#3.}}% Optional theorem note
\newtheoremstyle{blacknumbox} % Theorem style name
{0pt}% Space above
{0pt}% Space below
{\normalfont}% Body font
{}% Indent amount
{\small\bf\sffamily}% Theorem head font
{\;}% Punctuation after theorem head
{0.25em}% Space after theorem head
{\small\sffamily\thmname{#1}\nobreakspace\thmnumber{\@ifnotempty{#1}{}\@upn{#2}}% Theorem text (e.g. Theorem 2.1)
\thmnote{\nobreakspace\the\thm@notefont\sffamily\bfseries---\nobreakspace#3.}}% Optional theorem note
% Non-boxed/non-framed environments
\newtheoremstyle{ocrenum}% Theorem style name
{5pt}% Space above
{5pt}% Space below
{\normalfont}% Body font
{}% Indent amount
{\small\bf\sffamily\color{ocre}}% Theorem head font
{\;}% Punctuation after theorem head
{0.25em}% Space after theorem head
{\small\sffamily\color{ocre}\thmname{#1}\nobreakspace\thmnumber{\@ifnotempty{#1}{}\@upn{#2}}% Theorem text (e.g. Theorem 2.1)
\thmnote{\nobreakspace\the\thm@notefont\sffamily\bfseries\color{black}---\nobreakspace#3.}} % Optional theorem note
\makeatother
% Defines the theorem text style for each type of theorem to one of the three styles above
\newcounter{dummy}
\numberwithin{dummy}{section}
\theoremstyle{ocrenumbox}
\newtheorem{theoremeT}[dummy]{Theorem}
\newtheorem{problem}{Problem}[chapter]
\newtheorem{exerciseT}{Example}[chapter]
\theoremstyle{blacknumex}
\newtheorem{exampleT}{实例}[chapter]
\theoremstyle{blacknumbox}
\newtheorem{vocabulary}{Vocabulary}[chapter]
\newtheorem{definitionT}{定义}[section]
\newtheorem{corollaryT}[dummy]{Corollary}
\theoremstyle{ocrenum}
\newtheorem{proposition}[dummy]{Proposition}
%----------------------------------------------------------------------------------------
% DEFINITION OF COLORED BOXES
%----------------------------------------------------------------------------------------
\RequirePackage[framemethod=default]{mdframed} % Required for creating the theorem, definition, exercise and corollary boxes
% Theorem box
\newmdenv[skipabove=7pt,
skipbelow=7pt,
backgroundcolor=black!5,
linecolor=ocre,
innerleftmargin=5pt,
innerrightmargin=5pt,
innertopmargin=5pt,
leftmargin=0cm,
rightmargin=0cm,
innerbottommargin=5pt]{tBox}
% Exercise box
\newmdenv[skipabove=7pt,
skipbelow=7pt,
rightline=false,
leftline=true,
topline=false,
bottomline=false,
backgroundcolor=ocre!10,
linecolor=ocre,
innerleftmargin=5pt,
innerrightmargin=5pt,
innertopmargin=5pt,
innerbottommargin=5pt,
leftmargin=0cm,
rightmargin=0cm,
linewidth=4pt]{eBox}
% Definition box
\newmdenv[skipabove=7pt,
skipbelow=7pt,
rightline=false,
leftline=true,
topline=false,
bottomline=false,
linecolor=ocre,
innerleftmargin=5pt,
innerrightmargin=5pt,
innertopmargin=0pt,
leftmargin=0cm,
rightmargin=0cm,
linewidth=4pt,
innerbottommargin=0pt]{dBox}
% Corollary box
\newmdenv[skipabove=7pt,
skipbelow=7pt,
rightline=false,
leftline=true,
topline=false,
bottomline=false,
linecolor=gray,
backgroundcolor=black!5,
innerleftmargin=5pt,
innerrightmargin=5pt,
innertopmargin=5pt,
leftmargin=0cm,
rightmargin=0cm,
linewidth=4pt,
innerbottommargin=5pt]{cBox}
% Creates an environment for each type of theorem and assigns it a theorem text style from the "Theorem Styles" section above and a colored box from above
\newenvironment{theorem}{\begin{tBox}\begin{theoremeT}}{\end{theoremeT}\end{tBox}}
\newenvironment{exercise}{\begin{eBox}\begin{exerciseT}}{\hfill{\color{ocre}\tiny\ensuremath{\blacksquare}}\end{exerciseT}\end{eBox}}
\newenvironment{definition}{\begin{dBox}\begin{definitionT}}{\end{definitionT}\end{dBox}}
\newenvironment{example}{\begin{exampleT}}{\hfill{\tiny\ensuremath{\blacksquare}}\end{exampleT}}
\newenvironment{corollary}{\begin{cBox}\begin{corollaryT}}{\end{corollaryT}\end{cBox}}
%----------------------------------------------------------------------------------------
% REMARK ENVIRONMENT
%----------------------------------------------------------------------------------------
\newenvironment{remark}{\par\vspace{10pt}\small % Vertical white space above the remark and smaller font size
\begin{list}{}{
\leftmargin=35pt % Indentation on the left
\rightmargin=25pt}\item\ignorespaces % Indentation on the right
\makebox[-2.5pt]{\begin{tikzpicture}[overlay]
\node[draw=ocre!60,line width=1pt,circle,fill=ocre!25,font=\sffamily\bfseries,inner sep=2pt,outer sep=0pt] at (-15pt,0pt){\textcolor{ocre}{R}};\end{tikzpicture}} % Orange R in a circle
\advance\baselineskip 1pt}{\end{list}\vskip5pt} % Tighter line spacing and white space after remark
%----------------------------------------------------------------------------------------
% SECTION NUMBERING IN THE MARGIN
%----------------------------------------------------------------------------------------
%调整各级标题的段前段后间距
\makeatletter
\renewcommand{\@seccntformat}[1]{\llap{\textcolor{ocre}{\csname the#1\endcsname}\hspace{1em}}}
\renewcommand{\section}{\@startsection{section}{1}{\z@}
{-4ex \@plus -1ex \@minus -.4ex}
{1ex \@plus.2ex }
{\color{ublue}\normalfont\Large\sffamily\bfseries}}
\renewcommand{\subsection}{\@startsection {subsection}{2}{\z@}
{-3ex \@plus -0.1ex \@minus -.4ex}
{0.5ex \@plus.2ex }
{\normalfont\large\sffamily\bfseries}}
\renewcommand{\subsubsection}{\@startsection {subsubsection}{3}{\z@}
{-3ex \@plus -0.1ex \@minus -.4ex}
{.4ex \@plus.2ex }
{\normalfont\normalsize\sffamily\bfseries}}
\renewcommand\paragraph{\@startsection{paragraph}{4}{\z@}
{-2ex \@plus-.2ex \@minus .2ex}
{.1ex}
{\normalfont\small\sffamily\bfseries}}
%----------------------------------------------------------------------------------------
% PART HEADINGS
%----------------------------------------------------------------------------------------
% Numbered part in the table of contents
\newcommand{\@mypartnumtocformat}[2]{%
\setlength\fboxsep{0pt}%
\noindent\colorbox{ocre!20}{\strut\parbox[c][.7cm]{\ecart}{\color{ocre!70}\Large\sffamily\bfseries\centering#1}}\hskip\esp\colorbox{ocre!40}{\strut\parbox[c][.7cm]{\linewidth-\ecart-\esp}{\Large\sffamily\centering#2}}%
}
% Unnumbered part in the table of contents
\newcommand{\@myparttocformat}[1]{%
\setlength\fboxsep{0pt}%
\noindent\colorbox{ocre!40}{\strut\parbox[c][.7cm]{\linewidth}{\Large\sffamily\centering#1}}%
}
\newlength\esp
\setlength\esp{4pt}
\newlength\ecart
\setlength\ecart{1.2cm-\esp}
\newcommand{\thepartimage}{}%
\newcommand{\partimage}[1]{\renewcommand{\thepartimage}{#1}}%
\def\@part[#1]#2{%
\ifnum \c@secnumdepth >-2\relax%
\refstepcounter{part}%
\addcontentsline{toc}{part}{\texorpdfstring{\protect\@mypartnumtocformat{\thepart}{#1}}{\partname~\thepart\ ---\ #1}}
\else%
\addcontentsline{toc}{part}{\texorpdfstring{\protect\@myparttocformat{#1}}{#1}}%
\fi%
\startcontents%
\markboth{}{}%
{\thispagestyle{empty}%
\begin{tikzpicture}[remember picture,overlay]%
\node at (current page.north west){\begin{tikzpicture}[remember picture,overlay]%
\fill[ocre!20](0cm,0cm) rectangle (\paperwidth,-\paperheight);
\node[anchor=north] at (4cm,-3.25cm){\color{ocre!40}\fontsize{220}{100}\sffamily\bfseries\thepart};
\node[anchor=south east] at (\paperwidth-1cm,-\paperheight+1cm){\parbox[t][][t]{8.5cm}{
\printcontents{l}{0}{\setcounter{tocdepth}{1}}% The depth to which the Part mini table of contents displays headings; 0 for chapters only, 1 for chapters and sections and 2 for chapters, sections and subsections
}};
\node[anchor=north east] at (\paperwidth-1.5cm,-3.25cm){\parbox[t][][t]{15cm}{\strut\raggedleft\color{white}\fontsize{30}{30}\sffamily\bfseries#2}};
\end{tikzpicture}};
\end{tikzpicture}}%
\@endpart}
\def\@spart#1{%
\startcontents%
\phantomsection
{\thispagestyle{empty}%
\begin{tikzpicture}[remember picture,overlay]%
\node at (current page.north west){\begin{tikzpicture}[remember picture,overlay]%
\fill[ocre!20](0cm,0cm) rectangle (\paperwidth,-\paperheight);
\node[anchor=north east] at (\paperwidth-1.5cm,-3.25cm){\parbox[t][][t]{15cm}{\strut\raggedleft\color{white}\fontsize{30}{30}\sffamily\bfseries#1}};
\end{tikzpicture}};
\end{tikzpicture}}
\addcontentsline{toc}{part}{\texorpdfstring{%
\setlength\fboxsep{0pt}%
\noindent\protect\colorbox{ocre!40}{\strut\protect\parbox[c][.7cm]{\linewidth}{\Large\sffamily\protect\centering #1\quad\mbox{}}}}{#1}}%
\@endpart}
\def\@endpart{\vfil\newpage
\if@twoside
\if@openright
\null
\thispagestyle{empty}%
\newpage
\fi
\fi
\if@tempswa
\twocolumn
\fi}
%----------------------------------------------------------------------------------------
% SPECIAL FONTS
%----------------------------------------------------------------------------------------
\newcommand\bfnew[1]{\sffamily\bfseries{#1}}
%----------------------------------------------------------------------------------------
% CHAPTER HEADINGS
%----------------------------------------------------------------------------------------
% A switch to conditionally include a picture, implemented by Christian Hupfer
\newif\ifusechapterimage
\usechapterimagetrue
\newcommand{\thechapterimage}{}%
\newcommand{\chapterimage}[1]{\ifusechapterimage\renewcommand{\thechapterimage}{#1}\fi}%
\newcommand{\autodot}{.}
\def\@makechapterhead#1{%
{\parindent \z@ \raggedright \normalfont
\ifnum \c@secnumdepth >\m@ne
\if@mainmatter
\begin{tikzpicture}[remember picture,overlay]
\node at (current page.north west)
{\begin{tikzpicture}[remember picture,overlay]
\node[anchor=north west,inner sep=0pt] at (0,0) {\ifusechapterimage\includegraphics[width=\paperwidth]{\thechapterimage}\fi};
\draw[anchor=west] (\Gm@lmargin,-7cm) node [line width=2pt,rounded corners=15pt,draw=ocre,fill=white,fill opacity=0.5,inner sep=15pt]{\strut\makebox[22cm]{}};
\draw[anchor=west] (\Gm@lmargin+.3cm,-7cm) node {\huge\sffamily\bfseries\color{black}\thechapter\autodot~#1\strut};
\end{tikzpicture}};
\end{tikzpicture}
\else
\begin{tikzpicture}[remember picture,overlay]
\node at (current page.north west)
{\begin{tikzpicture}[remember picture,overlay]
\node[anchor=north west,inner sep=0pt] at (0,0) {\ifusechapterimage\includegraphics[width=\paperwidth]{\thechapterimage}\fi};
\draw[anchor=west] (\Gm@lmargin,-7cm) node [line width=2pt,rounded corners=15pt,draw=ocre,fill=white,fill opacity=0.5,inner sep=15pt]{\strut\makebox[22cm]{}};
\draw[anchor=west] (\Gm@lmargin+.3cm,-7cm) node {\huge\sffamily\bfseries\color{black}#1\strut};
\end{tikzpicture}};
\end{tikzpicture}
\fi\fi\par\vspace*{270\p@}}}
%-------------------------------------------
\def\@makeschapterhead#1{%
\begin{tikzpicture}[remember picture,overlay]
\node at (current page.north west)
{\begin{tikzpicture}[remember picture,overlay]
\node[anchor=north west,inner sep=0pt] at (0,0) {\ifusechapterimage\includegraphics[width=\paperwidth]{\thechapterimage}\fi};
\draw[anchor=west] (\Gm@lmargin,-7cm) node [line width=2pt,rounded corners=15pt,draw=ocre,fill=white,fill opacity=0.5,inner sep=15pt]{\strut\makebox[22cm]{}};
\draw[anchor=west] (\Gm@lmargin+.3cm,-7cm) node {\huge\sffamily\bfseries\color{black}#1\strut};
\end{tikzpicture}};
\end{tikzpicture}
\par\vspace*{270\p@}}
\makeatother
%----------------------------------------------------------------------------------------
% LINKS
%----------------------------------------------------------------------------------------
\usepackage{hyperref}
\hypersetup{hidelinks,backref=true,pagebackref=true,hyperindex=true,colorlinks=false,breaklinks=true,urlcolor=ocre,bookmarks=true,bookmarksopen=true}
%backref反向引用
%pagebackref反向引用页码
%hyperindex索引链接
%colorlinks彩色链接
%breaklinks允许链接断行
%urlcolor网页与电邮链接颜色
%bookmarks生成书签
%bookmarksopen书签目录展开
\usepackage{bookmark}
\bookmarksetup{
open,
numbered,
depth=2, %设置PDF的书签级别,2显示到subsection,3显示到subsubsection
addtohook={%
\ifnum\bookmarkget{level}=0 % chapter
\bookmarksetup{bold}%
\fi
\ifnum\bookmarkget{level}=-1 % part
\bookmarksetup{color=ocre,bold}%
\fi
}
}
%----------------------------------------------------------------------------------------
% NEW PAGE FOR SUBSECTION
%----------------------------------------------------------------------------------------
%\newcommand{\sectionnewpage}{\clearpage}
\newcommand{\sectionnewpage}{}
%----------------------------------------------------------------------------------------
% Chapter 3
%----------------------------------------------------------------------------------------
\usepackage{tikz}
\usetikzlibrary{arrows,decorations.pathreplacing}
\usetikzlibrary{positioning,fit,calc}
\usetikzlibrary{shadows} % LATEX and plain TEX when using Tik Z
\usetikzlibrary{mindmap,backgrounds} % mind map
\usepackage{type1cm}%设置公式字体
\usepackage{caption}%设置图片标题字体大小
\captionsetup{font={footnotesize}}
\usepackage{pstricks}
\DeclareMathOperator*{\argmax}{arg\,max}
\DeclareMathOperator*{\argmin}{arg\,min}
\usepackage{setspace}%调整行间距
%\usepackage{tocbibind}
%----------------------------------------------------------------------------------------
% Chapter 1
%----------------------------------------------------------------------------------------
\usepackage{chngpage}
\usepackage[justification=centering]{caption}%强制图片居中
\usepackage{subfigure}
\newcommand{\parinterval}{\noindent\hspace{2em}}%定义变量替代原来开头的控制缩进
\usepackage{tikz-qtree}
\usepackage{array}
\usepackage{booktabs}
\usepackage{bm}
\usetikzlibrary{shapes.misc}
\usepackage{appendix}
\usepackage{pgfplots}
\usepackage{tikz}
%----------------------------------------------------------------------------------------
% Chapter 4
%----------------------------------------------------------------------------------------
\usepackage{pgffor}%图片中使用\foreach语句
%\usepackage{ulem}%使用/sout
%----------------------------------------------------------------------------------------
% Chapter 6
%----------------------------------------------------------------------------------------
\usepackage{multirow}
\usepackage{tcolorbox}
\newcommand{\dash}{\raisebox{0.5mm}{------}}%中文破折号
\usepackage{colortbl} %table上色
\newlength{\base}
\newdimen\XCoord
\newdimen\YCoord
\newdimen\TMP
\newcommand*{\ExtractCoordinate}[1]{\path (#1); \pgfgetlastxy{\XCoord}{\YCoord};}%
\newcommand*{\ExtractX}[1]{\path (#1); \pgfgetlastxy{\XCoord}{\TMP};}%
\newcommand*{\ExtractY}[1]{\path (#1); \pgfgetlastxy{\TMP}{\YCoord};}%
\newcommand{\specialcell}[3][c]{%
\begin{tabular}[#1]{@{}#2@{}}#3\end{tabular}}
\usetikzlibrary{calc,intersections}
\usetikzlibrary{matrix}
\usetikzlibrary{patterns}
\usetikzlibrary{shadows.blur}
\usepgflibrary{arrows}
%\usetikzlibrary{arrows}
%\usetikzlibrary{decorations}
\usetikzlibrary{arrows,shapes}
%%%%%%%%%%%chapter5图片等---------------------------------------
\usepackage{tikz-3dplot}
\usepackage{pifont}
\tcbuselibrary{skins}
\definecolor{ublue}{rgb}{0.152,0.250,0.545}
\definecolor{ugreen}{rgb}{0,0.5,0}
\definecolor{lgreen}{rgb}{0.9,1,0.8}
\definecolor{xtgreen1}{rgb}{0.824,0.898,0.8}
\definecolor{xtgreen}{rgb}{0.914,0.945,0.902}
\definecolor{lightgray}{gray}{0.85}
%%%%%%%%%%%%appendix-------------------------------
\makeatletter
\def\UrlAlphabet{%
\do\a\do\b\do\c\do\d\do\e\do\f\do\g\do\h\do\i\do\j%
\do\k\do\l\do\m\do\n\do\o\do\p\do\q\do\r\do\s\do\t%
\do\u\do\v\do\w\do\x\do\y\do\z\do\A\do\B\do\C\do\D%
\do\E\do\F\do\G\do\H\do\I\do\J\do\K\do\L\do\M\do\N%
\do\O\do\P\do\Q\do\R\do\S\do\T\do\U\do\V\do\W\do\X%
\do\Y\do\Z}
\def\UrlDigits{\do\1\do\2\do\3\do\4\do\5\do\6\do\7\do\8\do\9\do\0}
\g@addto@macro{\UrlBreaks}{\UrlOrds}%特殊符号
\g@addto@macro{\UrlBreaks}{\UrlAlphabet}%26个字母表
\g@addto@macro{\UrlBreaks}{\UrlDigits}%10个阿拉伯数字
\makeatother
%上述设置的作用是URL自动换行
%%%%%%%%%%%chapter 7---------------------------------------
%\definecolor{myblack}{rgb}{0.15,0.15,0.15}
\definecolor{myblack}{rgb}{0.2,0.2,205.2}
\newlength{\hseg}
\newlength{\wnode}
\newlength{\hnode}
\newlength{\wseg}
\usepackage{collcell}
\usepackage[mathscr]{euscript}
\newcommand{\mychapter}[1]{第\ref{#1}章}%chapter用
\newcommand{\mysection}[1]{第\ref{#1}节}%section、subsection、subsubsection用
\ No newline at end of file
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论