Commit 13e75a8b by 曹润柘

update chapter5

parent 36421070
\relax
\providecommand\zref@newlabel[2]{}
\providecommand\hyper@newdestlabel[2]{}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {chapter}{\numberline {1}基于词的机器翻译建模}{7}{chapter.1}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\addvspace {10\p@ }}
\@writefile{lot}{\defcounter {refsection}{0}\relax }\@writefile{lot}{\addvspace {10\p@ }}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.1}词在翻译中的作用}{7}{section.1.1}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.1}{\ignorespaces 汉语到英语翻译实例及两种语言单词之间的对应关系\relax }}{8}{figure.caption.2}\protected@file@percent }
\providecommand*\caption@xref[2]{\@setref\relax\@undefined{#1}}
\newlabel{fig:5-1}{{1.1}{8}{汉语到英语翻译实例及两种语言单词之间的对应关系\relax }{figure.caption.2}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.2}{\ignorespaces 翻译过程中的分析、转换和生成\relax }}{8}{figure.caption.3}\protected@file@percent }
\newlabel{fig:5-2}{{1.2}{8}{翻译过程中的分析、转换和生成\relax }{figure.caption.3}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.2}一个简单实例}{9}{section.1.2}\protected@file@percent }
\newlabel{sec:simple-mt-example}{{1.2}{9}{一个简单实例}{section.1.2}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.2.1}如何进行翻译?}{9}{subsection.1.2.1}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.3}{\ignorespaces 人工翻译的过程\relax }}{9}{figure.caption.5}\protected@file@percent }
\newlabel{fig:5-3}{{1.3}{9}{人工翻译的过程\relax }{figure.caption.5}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{机器翻译流程}{10}{section*.6}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.4}{\ignorespaces 机器翻译的过程\ \raisebox {0.5mm}{------}\ 把单词的译文进行拼装,并找到最优的拼装路径\relax }}{11}{figure.caption.7}\protected@file@percent }
\newlabel{fig:5-4}{{1.4}{11}{机器翻译的过程\ \dash \ 把单词的译文进行拼装,并找到最优的拼装路径\relax }{figure.caption.7}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{人工翻译 vs. 机器翻译}{11}{section*.8}\protected@file@percent }
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.2.2}基本框架}{11}{subsection.1.2.2}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.5}{\ignorespaces 简单的统计机器翻译流程\relax }}{12}{figure.caption.9}\protected@file@percent }
\newlabel{fig:5-5}{{1.5}{12}{简单的统计机器翻译流程\relax }{figure.caption.9}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.2.3}单词翻译概率}{12}{subsection.1.2.3}\protected@file@percent }
\newlabel{chapter5.2.3}{{1.2.3}{12}{单词翻译概率}{subsection.1.2.3}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{什么是单词翻译概率?}{12}{section*.10}\protected@file@percent }
\@writefile{lot}{\defcounter {refsection}{0}\relax }\@writefile{lot}{\contentsline {table}{\numberline {1.1}{\ignorespaces 汉译英单词翻译概率\relax }}{13}{table.caption.11}\protected@file@percent }
\newlabel{tab:5-1}{{1.1}{13}{汉译英单词翻译概率\relax }{table.caption.11}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{如何从一个双语平行数据中学习?}{13}{section*.12}\protected@file@percent }
\newlabel{eq:5-1}{{1.1}{13}{如何从一个双语平行数据中学习?}{equation.1.2.1}{}}
\newlabel{eg:5-1}{{1.1}{13}{}{exampleT.1.1}{}}
\newlabel{eq:5-2}{{1.2}{14}{如何从一个双语平行数据中学习?}{equation.1.2.2}{}}
\newlabel{eq:5-3}{{1.4}{14}{如何从一个双语平行数据中学习?}{equation.1.2.3}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{如何从大量的双语平行数据中学习?}{14}{section*.13}\protected@file@percent }
\newlabel{eq:5-4}{{1.5}{14}{如何从大量的双语平行数据中学习?}{equation.1.2.5}{}}
\newlabel{eg:5-2}{{1.2}{14}{}{exampleT.1.2}{}}
\newlabel{eq:5-5}{{1.6}{15}{如何从大量的双语平行数据中学习?}{equation.1.2.6}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.2.4}句子级翻译模型}{15}{subsection.1.2.4}\protected@file@percent }
\newlabel{sec:sentence-level-translation}{{1.2.4}{15}{句子级翻译模型}{subsection.1.2.4}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.6}{\ignorespaces $\textrm {P}(\mathbf {t}|\mathbf {s})$在句子级翻译中的作用\relax }}{15}{figure.caption.14}\protected@file@percent }
\newlabel{fig:5-6}{{1.6}{15}{$\textrm {P}(\mathbf {t}|\mathbf {s})$在句子级翻译中的作用\relax }{figure.caption.14}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{基础模型}{15}{section*.15}\protected@file@percent }
\newlabel{eq:5-6}{{1.7}{15}{基础模型}{equation.1.2.7}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.7}{\ignorespaces 汉英互译句对及词对齐连接(蓝色虚线)\relax }}{16}{figure.caption.16}\protected@file@percent }
\newlabel{fig:5-7}{{1.7}{16}{汉英互译句对及词对齐连接(蓝色虚线)\relax }{figure.caption.16}{}}
\newlabel{eq:5-7}{{1.8}{17}{基础模型}{equation.1.2.8}{}}
\newlabel{eq:5-8}{{1.9}{17}{基础模型}{equation.1.2.9}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{生成流畅的译文}{17}{section*.17}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.8}{\ignorespaces 同一个源语言句子的不同译文所对应的$g(\cdot )$得分\relax }}{17}{figure.caption.18}\protected@file@percent }
\newlabel{fig:5-8}{{1.8}{17}{同一个源语言句子的不同译文所对应的$g(\cdot )$得分\relax }{figure.caption.18}{}}
\newlabel{eq:5-9}{{1.10}{18}{生成流畅的译文}{equation.1.2.10}{}}
\newlabel{eq:5-10}{{1.11}{18}{生成流畅的译文}{equation.1.2.11}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.9}{\ignorespaces 同一个源语言句子的不同译文所对应的语言模型得分和翻译模型得分\relax }}{18}{figure.caption.19}\protected@file@percent }
\newlabel{fig:5-9}{{1.9}{18}{同一个源语言句子的不同译文所对应的语言模型得分和翻译模型得分\relax }{figure.caption.19}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.2.5}解码}{19}{subsection.1.2.5}\protected@file@percent }
\newlabel{sec:simple-decoding}{{1.2.5}{19}{解码}{subsection.1.2.5}{}}
\newlabel{eq:5-11}{{1.12}{19}{解码}{equation.1.2.12}{}}
\newlabel{eq:5-12}{{1.13}{19}{解码}{equation.1.2.13}{}}
\newlabel{eq:5-13}{{1.14}{19}{解码}{equation.1.2.14}{}}
\@writefile{lot}{\defcounter {refsection}{0}\relax }\@writefile{lot}{\contentsline {table}{\numberline {1.2}{\ignorespaces 机器翻译搜索空间大小的示例\relax }}{19}{table.caption.20}\protected@file@percent }
\newlabel{tab:5-2}{{1.2}{19}{机器翻译搜索空间大小的示例\relax }{table.caption.20}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.10}{\ignorespaces 贪婪的机器翻译解码算法的伪代码\relax }}{20}{figure.caption.21}\protected@file@percent }
\newlabel{fig:5-10}{{1.10}{20}{贪婪的机器翻译解码算法的伪代码\relax }{figure.caption.21}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.11}{\ignorespaces 贪婪的机器翻译解码过程实例\relax }}{21}{figure.caption.23}\protected@file@percent }
\newlabel{fig:5-11}{{1.11}{21}{贪婪的机器翻译解码过程实例\relax }{figure.caption.23}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.3}噪声信道模型}{21}{section.1.3}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.12}{\ignorespaces 不同翻译候选空间的对比:人(左)vs 机器翻译 (右)\relax }}{22}{figure.caption.24}\protected@file@percent }
\newlabel{fig:5-12}{{1.12}{22}{不同翻译候选空间的对比:人(左)vs 机器翻译 (右)\relax }{figure.caption.24}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.13}{\ignorespaces 噪声信道模型,其中$\mathbf {s}$表示信宿,$\mathbf {t}$表示信源\relax }}{22}{figure.caption.25}\protected@file@percent }
\newlabel{fig:5-13}{{1.13}{22}{噪声信道模型,其中$\mathbf {s}$表示信宿,$\mathbf {t}$表示信源\relax }{figure.caption.25}{}}
\newlabel{eq:5-14}{{1.15}{22}{噪声信道模型}{equation.1.3.15}{}}
\newlabel{eq:5-15}{{1.16}{22}{噪声信道模型}{equation.1.3.16}{}}
\newlabel{eq:5-16}{{1.17}{23}{噪声信道模型}{equation.1.3.17}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.4}统计机器翻译的三个基本问题}{24}{section.1.4}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.14}{\ignorespaces IBM模型与公式\ref {eq:5-10}的对应关系\relax }}{24}{figure.caption.26}\protected@file@percent }
\newlabel{fig:5-14}{{1.14}{24}{IBM模型与公式\ref {eq:5-10}的对应关系\relax }{figure.caption.26}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.4.1}词对齐}{24}{subsection.1.4.1}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.15}{\ignorespaces 不同词对齐对比\relax }}{25}{figure.caption.27}\protected@file@percent }
\newlabel{fig:5-15}{{1.15}{25}{不同词对齐对比\relax }{figure.caption.27}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.16}{\ignorespaces 词对齐实例(``在''对应到$t_0$\relax }}{25}{figure.caption.28}\protected@file@percent }
\newlabel{fig:5-16}{{1.16}{25}{词对齐实例(``在''对应到$t_0$\relax }{figure.caption.28}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.4.2}基于词对齐的翻译模型}{25}{subsection.1.4.2}\protected@file@percent }
\newlabel{eq:5-17}{{1.18}{25}{基于词对齐的翻译模型}{equation.1.4.18}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.17}{\ignorespaces 一个汉译英句对的所有词对齐可能\relax }}{26}{figure.caption.29}\protected@file@percent }
\newlabel{fig:5-17}{{1.17}{26}{一个汉译英句对的所有词对齐可能\relax }{figure.caption.29}{}}
\newlabel{eq:5-18}{{1.19}{26}{基于词对齐的翻译模型}{equation.1.4.19}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.4.3}基于词对齐的翻译实例}{27}{subsection.1.4.3}\protected@file@percent }
\newlabel{eq:5-19}{{1.20}{27}{基于词对齐的翻译实例}{equation.1.4.20}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.5}IBM模型1-2}{27}{section.1.5}\protected@file@percent }
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.5.1}IBM模型1}{27}{subsection.1.5.1}\protected@file@percent }
\newlabel{eq:5-20}{{1.21}{28}{IBM模型1}{equation.1.5.21}{}}
\newlabel{eq:5-21}{{1.22}{28}{IBM模型1}{equation.1.5.22}{}}
\newlabel{eq:5-22}{{1.23}{28}{IBM模型1}{equation.1.5.23}{}}
\newlabel{eq:5-23}{{1.24}{28}{IBM模型1}{equation.1.5.24}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.18}{\ignorespaces 汉译英双语句对及词对齐\relax }}{28}{figure.caption.30}\protected@file@percent }
\newlabel{fig:5-18}{{1.18}{28}{汉译英双语句对及词对齐\relax }{figure.caption.30}{}}
\newlabel{eq:5-24}{{1.25}{29}{IBM模型1}{equation.1.5.25}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.19}{\ignorespaces 公式{\ref {eq:5-24}}第一部分实例\relax }}{29}{figure.caption.31}\protected@file@percent }
\newlabel{fig:5-19}{{1.19}{29}{公式{\ref {eq:5-24}}第一部分实例\relax }{figure.caption.31}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.5.2}IBM模型2}{29}{subsection.1.5.2}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.20}{\ignorespaces 不同的译文导致不同IBM模型1得分的情况\relax }}{29}{figure.caption.32}\protected@file@percent }
\newlabel{fig:5-20}{{1.20}{29}{不同的译文导致不同IBM模型1得分的情况\relax }{figure.caption.32}{}}
\newlabel{eq:5-25}{{1.26}{30}{IBM模型2}{equation.1.5.26}{}}
\newlabel{eq:5-26}{{1.27}{30}{IBM模型2}{equation.1.5.27}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.5.3}解码及计算优化}{30}{subsection.1.5.3}\protected@file@percent }
\newlabel{eq:5-27}{{1.28}{30}{解码及计算优化}{equation.1.5.28}{}}
\newlabel{eq:5-28}{{1.29}{31}{解码及计算优化}{equation.1.5.29}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.21}{\ignorespaces $\sum \limits _{a_1=0}^{l} ... \sum \limits _{a_m=0}^{l} \prod \limits _{j=1}^{m} f(s_j|t_{a_j}) = \prod \limits _{j=1}^{m} \sum \limits _{i=0}^{l} f(s_j|t_i)$的实例\relax }}{31}{figure.caption.33}\protected@file@percent }
\newlabel{fig:5-21}{{1.21}{31}{$\sum \limits _{a_1=0}^{l} ... \sum \limits _{a_m=0}^{l} \prod \limits _{j=1}^{m} f(s_j|t_{a_j}) = \prod \limits _{j=1}^{m} \sum \limits _{i=0}^{l} f(s_j|t_i)$的实例\relax }{figure.caption.33}{}}
\newlabel{eq:5-64}{{1.30}{31}{解码及计算优化}{equation.1.5.30}{}}
\newlabel{eq:5-65}{{1.31}{31}{解码及计算优化}{equation.1.5.30}{}}
\newlabel{eq:5-29}{{1.31}{31}{解码及计算优化}{equation.1.5.30}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.5.4}训练}{31}{subsection.1.5.4}\protected@file@percent }
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{目标函数}{32}{section*.34}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.22}{\ignorespaces 一个目标函数的最优解\relax }}{32}{figure.caption.35}\protected@file@percent }
\newlabel{fig:5-22}{{1.22}{32}{一个目标函数的最优解\relax }{figure.caption.35}{}}
\newlabel{eq:5-30}{{1.32}{32}{目标函数}{equation.1.5.32}{}}
\newlabel{eq:5-31}{{1.33}{32}{目标函数}{equation.1.5.33}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsubsection}{优化}{33}{section*.36}\protected@file@percent }
\newlabel{eq:5-32}{{1.33}{33}{优化}{equation.1.5.33}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.23}{\ignorespaces 拉格朗日乘数法(IBM模型1)\relax }}{33}{figure.caption.37}\protected@file@percent }
\newlabel{fig:5-23}{{1.23}{33}{拉格朗日乘数法(IBM模型1)\relax }{figure.caption.37}{}}
\newlabel{eq:5-33}{{1.34}{33}{优化}{equation.1.5.34}{}}
\newlabel{eq:5-34}{{1.35}{34}{优化}{equation.1.5.35}{}}
\newlabel{eq:5-35}{{1.36}{34}{优化}{equation.1.5.36}{}}
\newlabel{eq:5-36}{{1.37}{34}{优化}{equation.1.5.37}{}}
\newlabel{eq:5-37}{{1.38}{34}{优化}{equation.1.5.38}{}}
\newlabel{eq:5-38}{{1.39}{35}{优化}{equation.1.5.39}{}}
\newlabel{eq:5-39}{{1.40}{35}{优化}{equation.1.5.40}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.24}{\ignorespaces IBM模型迭代过程示意图\relax }}{35}{figure.caption.38}\protected@file@percent }
\newlabel{fig:5-24}{{1.24}{35}{IBM模型迭代过程示意图\relax }{figure.caption.38}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.25}{\ignorespaces 公式\ref {eq:5-39}的解释\relax }}{35}{figure.caption.39}\protected@file@percent }
\newlabel{fig:5-25}{{1.25}{35}{公式\ref {eq:5-39}的解释\relax }{figure.caption.39}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.26}{\ignorespaces 频次(左)和期望频次(右)的计算过程\relax }}{36}{figure.caption.40}\protected@file@percent }
\newlabel{fig:5-26}{{1.26}{36}{频次(左)和期望频次(右)的计算过程\relax }{figure.caption.40}{}}
\newlabel{eq:5-40}{{1.42}{36}{优化}{equation.1.5.42}{}}
\newlabel{eq:5-41}{{1.43}{36}{优化}{equation.1.5.43}{}}
\newlabel{eq:5-42}{{1.44}{36}{优化}{equation.1.5.44}{}}
\newlabel{eq:5-43}{{1.45}{36}{优化}{equation.1.5.45}{}}
\newlabel{eq:5-44}{{1.46}{36}{优化}{equation.1.5.46}{}}
\newlabel{eq:5-45}{{1.47}{37}{优化}{equation.1.5.47}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.27}{\ignorespaces $f(s_u|t_v)$的计算公式和迭代过程\relax }}{37}{figure.caption.41}\protected@file@percent }
\newlabel{fig:5-27}{{1.27}{37}{$f(s_u|t_v)$的计算公式和迭代过程\relax }{figure.caption.41}{}}
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.28}{\ignorespaces EM算法流程图(IBM模型1)\relax }}{37}{figure.caption.42}\protected@file@percent }
\newlabel{fig:5-28}{{1.28}{37}{EM算法流程图(IBM模型1)\relax }{figure.caption.42}{}}
\newlabel{eq:5-46}{{1.48}{37}{优化}{equation.1.5.48}{}}
\newlabel{eq:5-47}{{1.50}{38}{优化}{equation.1.5.49}{}}
\newlabel{eq:5-48}{{1.52}{38}{优化}{equation.1.5.51}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.6}隐马尔可夫模型}{38}{section.1.6}\protected@file@percent }
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.6.1}隐马尔可夫模型}{38}{subsection.1.6.1}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.29}{\ignorespaces 抛硬币的隐马尔可夫模型实例\relax }}{39}{figure.caption.43}\protected@file@percent }
\newlabel{fig:5-29}{{1.29}{39}{抛硬币的隐马尔可夫模型实例\relax }{figure.caption.43}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {subsection}{\numberline {1.6.2}词对齐模型}{39}{subsection.1.6.2}\protected@file@percent }
\@writefile{lof}{\defcounter {refsection}{0}\relax }\@writefile{lof}{\contentsline {figure}{\numberline {1.30}{\ignorespaces 汉译英句对及对齐\relax }}{40}{figure.caption.44}\protected@file@percent }
\newlabel{fig:5-30}{{1.30}{40}{汉译英句对及对齐\relax }{figure.caption.44}{}}
\newlabel{eq:5-49}{{1.53}{40}{词对齐模型}{equation.1.6.53}{}}
\newlabel{eq:5-50}{{1.54}{40}{词对齐模型}{equation.1.6.54}{}}
\newlabel{eq:5-51}{{1.55}{40}{词对齐模型}{equation.1.6.55}{}}
\@writefile{toc}{\defcounter {refsection}{0}\relax }\@writefile{toc}{\contentsline {section}{\numberline {1.7}小结及深入阅读}{40}{section.1.7}\protected@file@percent }
\@setckpt{Chapter5/chapter5}{
\setcounter{page}{42}
\setcounter{equation}{55}
\setcounter{enumi}{0}
\setcounter{enumii}{0}
\setcounter{enumiii}{0}
\setcounter{enumiv}{0}
\setcounter{footnote}{6}
\setcounter{mpfootnote}{0}
\setcounter{part}{0}
\setcounter{chapter}{1}
\setcounter{section}{7}
\setcounter{subsection}{0}
\setcounter{subsubsection}{0}
\setcounter{paragraph}{0}
\setcounter{subparagraph}{0}
\setcounter{figure}{30}
\setcounter{table}{2}
\setcounter{tabx@nest}{0}
\setcounter{listtotal}{0}
\setcounter{listcount}{0}
\setcounter{liststart}{0}
\setcounter{liststop}{0}
\setcounter{citecount}{0}
\setcounter{citetotal}{0}
\setcounter{multicitecount}{0}
\setcounter{multicitetotal}{0}
\setcounter{instcount}{13}
\setcounter{maxnames}{3}
\setcounter{minnames}{1}
\setcounter{maxitems}{3}
\setcounter{minitems}{1}
\setcounter{citecounter}{0}
\setcounter{maxcitecounter}{0}
\setcounter{savedcitecounter}{0}
\setcounter{uniquelist}{0}
\setcounter{uniquename}{0}
\setcounter{refsection}{0}
\setcounter{refsegment}{0}
\setcounter{maxextratitle}{0}
\setcounter{maxextratitleyear}{0}
\setcounter{maxextraname}{2}
\setcounter{maxextradate}{0}
\setcounter{maxextraalpha}{0}
\setcounter{abbrvpenalty}{50}
\setcounter{highnamepenalty}{50}
\setcounter{lownamepenalty}{25}
\setcounter{maxparens}{3}
\setcounter{parenlevel}{0}
\setcounter{mincomprange}{10}
\setcounter{maxcomprange}{100000}
\setcounter{mincompwidth}{1}
\setcounter{afterword}{0}
\setcounter{savedafterword}{0}
\setcounter{annotator}{0}
\setcounter{savedannotator}{0}
\setcounter{author}{0}
\setcounter{savedauthor}{0}
\setcounter{bookauthor}{0}
\setcounter{savedbookauthor}{0}
\setcounter{commentator}{0}
\setcounter{savedcommentator}{0}
\setcounter{editor}{0}
\setcounter{savededitor}{0}
\setcounter{editora}{0}
\setcounter{savededitora}{0}
\setcounter{editorb}{0}
\setcounter{savededitorb}{0}
\setcounter{editorc}{0}
\setcounter{savededitorc}{0}
\setcounter{foreword}{0}
\setcounter{savedforeword}{0}
\setcounter{holder}{0}
\setcounter{savedholder}{0}
\setcounter{introduction}{0}
\setcounter{savedintroduction}{0}
\setcounter{namea}{0}
\setcounter{savednamea}{0}
\setcounter{nameb}{0}
\setcounter{savednameb}{0}
\setcounter{namec}{0}
\setcounter{savednamec}{0}
\setcounter{translator}{0}
\setcounter{savedtranslator}{0}
\setcounter{shortauthor}{0}
\setcounter{savedshortauthor}{0}
\setcounter{shorteditor}{0}
\setcounter{savedshorteditor}{0}
\setcounter{labelname}{0}
\setcounter{savedlabelname}{0}
\setcounter{institution}{0}
\setcounter{savedinstitution}{0}
\setcounter{lista}{0}
\setcounter{savedlista}{0}
\setcounter{listb}{0}
\setcounter{savedlistb}{0}
\setcounter{listc}{0}
\setcounter{savedlistc}{0}
\setcounter{listd}{0}
\setcounter{savedlistd}{0}
\setcounter{liste}{0}
\setcounter{savedliste}{0}
\setcounter{listf}{0}
\setcounter{savedlistf}{0}
\setcounter{location}{0}
\setcounter{savedlocation}{0}
\setcounter{organization}{0}
\setcounter{savedorganization}{0}
\setcounter{origlocation}{0}
\setcounter{savedoriglocation}{0}
\setcounter{origpublisher}{0}
\setcounter{savedorigpublisher}{0}
\setcounter{publisher}{0}
\setcounter{savedpublisher}{0}
\setcounter{language}{0}
\setcounter{savedlanguage}{0}
\setcounter{origlanguage}{0}
\setcounter{savedoriglanguage}{0}
\setcounter{pageref}{0}
\setcounter{savedpageref}{0}
\setcounter{textcitecount}{0}
\setcounter{textcitetotal}{0}
\setcounter{textcitemaxnames}{0}
\setcounter{biburlbigbreakpenalty}{100}
\setcounter{biburlbreakpenalty}{200}
\setcounter{biburlnumpenalty}{0}
\setcounter{biburlucpenalty}{0}
\setcounter{biburllcpenalty}{0}
\setcounter{smartand}{1}
\setcounter{bbx:relatedcount}{0}
\setcounter{bbx:relatedtotal}{0}
\setcounter{parentequation}{0}
\setcounter{notation}{0}
\setcounter{dummy}{0}
\setcounter{problem}{0}
\setcounter{exerciseT}{0}
\setcounter{exampleT}{2}
\setcounter{vocabulary}{0}
\setcounter{definitionT}{0}
\setcounter{mdf@globalstyle@cnt}{0}
\setcounter{mdfcountframes}{0}
\setcounter{mdf@env@i}{0}
\setcounter{mdf@env@ii}{0}
\setcounter{mdf@zref@counter}{0}
\setcounter{Item}{0}
\setcounter{Hfootnote}{6}
\setcounter{Hy@AnnotLevel}{0}
\setcounter{bookmark@seq@number}{0}
\setcounter{caption@flags}{0}
\setcounter{continuedfloat}{0}
\setcounter{cp@cnt}{0}
\setcounter{cp@tempcnt}{0}
\setcounter{subfigure}{0}
\setcounter{lofdepth}{1}
\setcounter{subtable}{0}
\setcounter{lotdepth}{1}
\setcounter{@pps}{0}
\setcounter{@ppsavesec}{0}
\setcounter{@ppsaveapp}{0}
\setcounter{tcbbreakpart}{0}
\setcounter{tcblayer}{0}
\setcounter{tcolorbox@number}{0}
\setcounter{section@level}{1}
}
......@@ -24,13 +24,13 @@
\chapter{基于词的机器翻译建模}
\parinterval 使用概率化的方法对翻译问题进行建模是机器翻译发展中的重要里程碑。这种思想也影响了当今的统计机器翻译和神经机器翻译方法。虽然技术不断发展,传统的统计模型已经不再``新鲜'',但它对于今天机器翻译的研究仍然有着重要的启示作用。在了解前沿、展望未来的同时,我们更要冷静的思考前人给我们带来了什么。基于此,本章将介绍统计机器翻译的开山之作\ \dash \ IBM模型,它提出了使用统计模型进行翻译的思想,并在建模中引入了单词对齐这一重要概念。IBM模型由Peter F. Brown等人于上世纪九十年代初提出\cite{Peter1993The}。客观的说,这项工作的视野和对问题的理解,已经超过当时很多人所能看到的东西,其衍生出来的一系列方法和新的问题还被后人花费将近10年的时间来进行研究与讨论。时至今日,IBM模型中的一些思想仍然影响着很多研究工作。
\parinterval {\color{red}使用概率化的方法对翻译问题进行建模是机器翻译发展中的重要里程碑。这种思想也影响了当今的统计机器翻译和神经机器翻译方法。虽然技术不断发展,传统的统计模型已经不再``新鲜'',但它对于今天机器翻译的研究仍然有着重要的启示作用。在了解前沿、展望未来的同时,我们更要冷静的思考前人给我们带来了什么。基于此,本章将介绍统计机器翻译的开山之作\ \dash \ IBM模型,它提出了使用统计模型进行翻译的思想,并在建模中引入了单词对齐这一重要概念。IBM模型由Peter F. Brown等人于上世纪九十年代初提出\cite{Peter1993The}。客观的说,这项工作的视野和对问题的理解,已经超过当时很多人所能看到的东西,其衍生出来的一系列方法和新的问题还被后人花费将近10年的时间来进行研究与讨论。时至今日,IBM模型中的一些思想仍然影响着很多研究工作。}
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
\section{什么是基于词的翻译模型}
\section{词在翻译中的作用}
\parinterval 在机器翻译中,我们希望得到一个源语言到目标语言的翻译。对于人类来说这个问题很简单,但是让计算机做这样的工作却很困难,因为我们需要把翻译``描述''成计算机可以计算的形式。这里面临的第一个问题是:如何对翻译进行建模?从计算机的角度来看,这就需要把自然语言的翻译问题转换为计算机可计算的问题。
......@@ -74,7 +74,7 @@
%----------------------------------------------------------------------------------------
\sectionnewpage
\section{构建一个简单的机器翻译系统}
\section{一个简单实例}
\label{sec:simple-mt-example}
\parinterval 本节首先对比人工翻译和机器翻译过程的异同点,从中归纳出构建机器翻译系统的两个主要步骤:训练和解码。之后,会从学习翻译知识和运用翻译知识两个方面描述如何构建一个简单的机器翻译系统。
......@@ -285,7 +285,7 @@ $\mathbf{t}^{[2]}$ = So\; ,\; what\; is\; human\; {\color{red}translation}\; ?
\label{eg:5-2}
\end{example}
\parinterval 举个例子来说明在多个句子上计算单词翻译概率的方法。例\ref{eg:5-2}展示了一个由两个句对构成的平行语料库。其中,$\mathbf{s}^{[1]}$$\mathbf{s}^{[2]}$分别表示第一个句对和第二个句对的源语言句子,$\mathbf{t}^{[1]}$$\mathbf{t}^{[2]}$表示对应的目标语言句子。于是,``翻译''和``translation''的翻译概率为
\parinterval 举个例子来说明在多个句子上计算单词翻译概率的方法。\ref{eg:5-2}展示了一个由两个句对构成的平行语料库。其中,$\mathbf{s}^{[1]}$$\mathbf{s}^{[2]}$分别表示第一个句对和第二个句对的源语言句子,$\mathbf{t}^{[1]}$$\mathbf{t}^{[2]}$表示对应的目标语言句子。于是,``翻译''和``translation''的翻译概率为
{\small
\begin{eqnarray}
{\textrm{P}(\textrm{``翻译''},\textrm{``translation''})} & = & {\frac{c(\textrm{``翻译''},\textrm{``translation''};\mathbf{s}^{[1]},\mathbf{t}^{[1]})+c(\textrm{``翻译''},\textrm{``translation''};\mathbf{s}^{[2]},\mathbf{t}^{[2]})}{\sum_{x',y'} c(x',y';\mathbf{s}^{[1]},\mathbf{t}^{[1]}) + \sum_{x',y'} c(x',y';\mathbf{s}^{[2]},\mathbf{t}^{[2]})}} \nonumber \\
......@@ -503,8 +503,6 @@ g(\mathbf{s},\mathbf{t}) \equiv \prod_{j,i \in \widehat{A}}{\textrm{P}(s_j,t_i)}
\vspace{-1.8em}
该算法的核心在于,系统一直维护一个当前最好的结果,之后每一步考虑扩展这个结果的所有可能,并计算模型得分,然后再保留扩展后的最好结果。注意,在每一步中,只有排名第一的结果才会被保留,其他结果都会被丢弃。这也体现了贪婪的思想。显然这个方法不能保证搜索到全局最优的结果,但是由于每次扩展只考虑一个最好的结果,因此该方法速度很快。图\ref{fig:5-11}给出了算法执行过程的简单示例。当然,机器翻译的解码方法有很多,这里仅仅使用简单的贪婪搜索方法来解决机器翻译的解码问题,在后续章节会对更加优秀的解码方法进行介绍。
\parinterval\ref{sec:simple-mt-example}节中,我们实现了一个简单的基于词的统计机器翻译模型,内容涉及建模、训练和解码。但是,还有很多问题还没有进行深入讨论,比如,如何处理空翻译?如何对调序问题进行建模?如何用更严密的数学模型描述翻译过程?如何对更加复杂的统计模型进行训练?等等。针对以上问题,本节将系统的介绍IBM统计机器翻译模型。作为经典的机器翻译模型,对IBM模型的学习将帮助我们建立对自然语言处理问题的系统化建模思想,特别是对问题的数学描述方法将会成为理解本书后续内容的基础工具。
%----------------------------------------------------------------------------------------
% NEW SECTION
%----------------------------------------------------------------------------------------
......@@ -512,6 +510,9 @@ g(\mathbf{s},\mathbf{t}) \equiv \prod_{j,i \in \widehat{A}}{\textrm{P}(s_j,t_i)}
\section{噪声信道模型}
\vspace{0.5em}
\parinterval\ref{sec:simple-mt-example}节中,我们实现了一个简单的基于词的统计机器翻译模型,内容涉及建模、训练和解码。但是,还有很多问题还没有进行深入讨论,比如,如何处理空翻译?如何对调序问题进行建模?如何用更严密的数学模型描述翻译过程?如何对更加复杂的统计模型进行训练?等等。针对以上问题,本节将系统的介绍IBM统计机器翻译模型。作为经典的机器翻译模型,对IBM模型的学习将帮助我们建立对自然语言处理问题的系统化建模思想,特别是对问题的数学描述方法将会成为理解本书后续内容的基础工具。
\parinterval 首先,重新思考一下人类进行翻译的过程。对于给定的源语句$\mathbf{s}$,人不会像计算机一样尝试很多的可能,而是快速准确的翻译出一个或者少数几个正确的译文。在人看来,除了正确的译文外,其他的翻译都是不正确的,或者说除了少数的译文人甚至都不会考虑太多其他的可能性。但是,在统计机器翻译的世界里,没有译文是不可能的。换句话说,对于源语言句子$\mathbf{s}$,所有目标语词串$\mathbf{t}$都是可能的译文,只是可能性大小不同。即每对$(\mathbf{s},\mathbf{t})$都有一个概率值$\textrm{P}(\mathbf{t}|\mathbf{s})$来描述$\mathbf{s}$翻译为$\mathbf{t}$的好与坏(图\ref{fig:5-12})。
%----------------------------------------------
......@@ -537,7 +538,7 @@ g(\mathbf{s},\mathbf{t}) \equiv \prod_{j,i \in \widehat{A}}{\textrm{P}(s_j,t_i)}
\end{figure}
%----------------------------------------------
\parinterval 举个例子对于汉译英的翻译任务,汉语句子$\mathbf{s}$可以被看作是英语句子$\mathbf{t}$加入噪声通过信道后得到的结果。换句话说,英语句子经过噪声-信道传输时发生了变化,在信道的输出端呈现为汉语句子。于是需要根据观察到的汉语特征,通过概率$\textrm{P}(\mathbf{t}|\mathbf{s})$猜测最为可能的英语句子。这个找到最可能的目标语句(信源)的过程也被称为
\parinterval 举个例子对于汉译英的翻译任务,汉语句子$\mathbf{s}$可以被看作是英语句子$\mathbf{t}$加入噪声通过信道后得到的结果。换句话说,英语句子经过噪声-信道传输时发生了变化,在信道的输出端呈现为汉语句子。于是需要根据观察到的汉语特征,通过概率$\textrm{P}(\mathbf{t}|\mathbf{s})$猜测最为可能的英语句子。这个找到最可能的目标语句(信源)的过程也被称为
{\small\sffamily\bfseries{解码}}(Decoding)。直到今天,解码这个概念也被广泛的使用在机器翻译及相关任务中。这个过程也可以表述为:给定输入$\mathbf{s}$,找到最可能的输出$\mathbf{t}$,使得$\textrm{P}(\mathbf{t}|\mathbf{s})$达到最大:
\begin{eqnarray}
\widehat{\mathbf{t}}=\argmax_{\mathbf{t}}\textrm{P}(\mathbf{t}|\mathbf{s})
......@@ -1113,7 +1114,9 @@ a(i|j,m,l) &=\frac{\sum_{k=0}^{K}c_{\mathbb{E}}(i|j;\mathbf{s}^{[k]},\mathbf{t}^
\section{隐马尔可夫模型}
\parinterval IBM模型可以得到双语句子间的词对齐,因此也有很多工作在这个模型的基础上对词对齐方法进行改进。其中一个比较有代表性的工作是基于隐马尔可夫模型的方法\cite{vogel1996hmm},它可以被看作是IBM 模型2的升级版本。{\color{red}这部分重点说HMM在机器翻译和对齐中的应用}
\parinterval {\color{red}IBM模型可以得到双语句子间的词对齐,因此也有很多工作在这个模型的基础上对词对齐方法进行改进。其中一个比较有代表性的工作是基于隐马尔可夫模型的方法\cite{vogel1996hmm},它可以被看作是IBM 模型2的升级版本。
这部分重点说HMM在机器翻译和对齐中的应用}
%----------------------------------------------------------------------------------------
% NEW SUB-SECTION
......@@ -1196,7 +1199,7 @@ a(i|j,m,l) &=\frac{\sum_{k=0}^{K}c_{\mathbb{E}}(i|j;\mathbf{s}^{[k]},\mathbf{t}^
\parinterval 本章对IBM系列模型进行了全面的介绍和讨论,从一个简单的基于单词的翻译模型开始,本章以建模、解码、训练多个维度对统计机器翻译进行了描述,期间也涉及了词对齐、优化等多个重要概念。IBM 模型共分为5个模型,对翻译问题的建模依次由浅入深,同时模型复杂度也依次增加。IBM模型作为入门统计机器翻译的``必经之路'',其思想对今天的机器翻译仍然产生着影响。虽然单独使用IBM模型进行机器翻译现在已经不多见,甚至很多从事神经机器翻译等前沿研究的人对IBM模型已经逐渐淡忘,但是不能否认IBM模型标志着一个时代的开始。从某种意义上,当使用公式$\hat{\mathbf{t}} = \argmax_{\mathbf{t}} \textrm{P}(\mathbf{t}|\mathbf{s})$描述机器翻译问题的时候,或多或少都在与IBM模型使用相似的思想。
{\color{red}词对齐}
{\color{red}词对齐需要补充进去}
\parinterval 当然,本书也无法涵盖IBM模型的所有内涵,很多内容需要感兴趣的读者继续研究和挖掘,有两个方向可以考虑:
......
Markdown 格式
0%
您添加了 0 到此讨论。请谨慎行事。
请先完成此评论的编辑!
注册 或者 后发表评论