Skip to content

M
mtbookv2
Commit b896e777 by 曹润柘
Options

合并分支 'caorunzhe' 到 'master' 

Caorunzhe

查看合并请求 !688
parents c42ca33d bb5e660a
正在显示 包含 305 行增加 和 7 行删除
Chapter17/Figures/figure-application-of-multimodal-machine-translation-to-multitask-learning.tex 0 → 100644
\tikzstyle{coder} = [rectangle,thick,rounded corners,minimum width=2.8cm,minimum height=1.3cm,text centered,draw=black,fill=red!25]
\begin{tikzpicture}[node distance = 0,scale = 1]
\tikzstyle{every node}=[scale=1]
\node(x)[]{\LARGE x};
\node(encoder)[coder, above of = x,yshift=2cm]{\Large{编码器}};
\node(decoder_left)[coder, above of = encoder, yshift=2.5cm,fill=blue!25]{\Large{解码器}};
\node(y_hat)[above of = decoder_left, yshift=2cm]{\LARGE{$\rm y'$}};
\node(y)[above of = decoder_left, xshift=-3cm]{\LARGE{$\rm y$}};
\node(decoder_right)[coder, above of = encoder, xshift=5cm,fill=yellow!25]{\Large{解码器}};
\node(figure)[draw=white,above of = decoder_right,yshift=2.5cm,scale=0.25] {\includegraphics[width=0.62\textwidth]{./Chapter17/Figures/figure-bank-without-attention.png}};
\draw[->,very thick](x)to(encoder);
\draw[->,very thick](encoder)to(decoder_left)node[right,xshift=-0.1cm,yshift=-1.25cm,scale=1.1]{翻译};
\draw[->,very thick](decoder_left)to(y_hat);
\draw[->,very thick](y)to(decoder_left);
\draw[->,very thick](encoder)to(decoder_right)node[right,xshift=-3.5cm,yshift=0.25cm,scale=1.1]{生成图片};
\draw[->,very thick](decoder_right)to(figure);
\end{tikzpicture}
\ No newline at end of file
Chapter17/Figures/figure-comparison-of-attention-mechanism-of-target-word-bank.tex 0 → 100644
\begin{tikzpicture}[node distance = 0,scale = 0.7]
\tikzstyle{every node}=[scale=0.7]
\node[draw=white] (input) at (0,0){\includegraphics[width=0.62\textwidth]{./Chapter17/Figures/figure-bank-without-attention.png}};(1.9,-1.4);
\node[draw=white] (input) at (10,0){\includegraphics[width=0.62\textwidth]{./Chapter17/Figures/figure-bank-with-attention.png}};(1.9,-1.4);
\end{tikzpicture}
\ No newline at end of file
Chapter17/Figures/figure-the-encoder-explicitly-incorporates-semantic-information.tex 0 → 100644
\tikzstyle{word} = [rectangle,thick,minimum width=2cm,minimum height=0.7cm,text centered,]
\begin{tikzpicture}[node distance = 0,scale = 0.9]
\tikzstyle{every node}=[scale=0.9]
\node(figure)[draw=white,scale=0.4] {\includegraphics[width=0.62\textwidth]{./Chapter17/Figures/figure-bank-without-attention.png}};
\node(river)[word, right of = figure, xshift=5cm, yshift=0.35cm, fill=blue!45]{river};
\node(mountain)[word, above of = river, yshift=0.75cm, fill=blue!45]{mountain};
\node(child)[word, above of = mountain, yshift=0.75cm, fill=blue!15]{child};
\node(man)[word, above of = child, yshift=0.75cm, fill=blue!25]{man};
\node(jump)[word, below of = river, yshift=-0.75cm, fill=blue!30]{jump};
\node(bank)[word, below of = jump, yshift=-0.75cm, fill=blue!65]{bank};
\node(sky)[word, below of = bank, yshift=-0.75cm, fill=blue!30]{sky};
\node(tree)[word, below of = sky, yshift=-0.75cm, fill=blue!15]{tree};
\node(cir)[circle,very thick, minimum width=0.6cm, xshift=8cm, draw=black]{};
\node(decoder)[rectangle, rounded corners, minimum width=2.5cm, minimum height=1.2cm, right of = cir,xshift=3cm, draw=black, fill=blue!25]{\large{解码器}};
\node(yn_1)[below of = decoder,yshift=-2cm,scale=1.2]{$\rm y_{n-1}$};
\node(yn_2)[above of = decoder,yshift=2cm,scale=1.2]{$\rm y'_{n-1}$(bank)};
\draw[->, very thick]([xshift=0.1cm]figure.east)to([xshift=2cm]figure.east);
\draw[-,very thick]([xshift=-0.03cm]cir.east)to([xshift=0.03cm]cir.west);
\draw[-,very thick]([yshift=0.03cm]cir.south)to([yshift=-0.03cm]cir.north);
\draw[->, very thick]([xshift=0.1cm]cir.east)to([xshift=-0.1cm]decoder.west);
\draw[->, very thick](yn_1)to([yshift=-0.1cm]decoder.south);
\draw[->, very thick]([yshift=0.1cm]decoder.north)to(yn_2);
\draw[->, thick, color=blue!45]([xshift=0.05cm]river.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!45]([xshift=0.05cm]mountain.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!15]([xshift=0.05cm]child.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!25]([xshift=0.05cm]man.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!30]([xshift=0.05cm]jump.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!65]([xshift=0.05cm]bank.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!30]([xshift=0.05cm]sky.east)to([xshift=-0.05cm]cir.west);
\draw[->, thick, color=blue!15]([xshift=0.05cm]tree.east)to([xshift=-0.05cm]cir.west);
\end{tikzpicture}
\ No newline at end of file
Chapter17/Figures/figure-traditional-methods-of-image-description.tex 0 → 100644
\definecolor{color_gray}{rgb}{0.278,0.337,0.426}
\definecolor{color_green}{rgb}{0.663,0.82,0.557}
\definecolor{color_orange}{rgb}{0.957,0.694,0.514}
\definecolor{color_blue}{rgb}{0.335,0.708,0.735}
\tikzstyle{description} = [rectangle,rounded corners=1mm, minimum width=3cm,minimum height=0.6cm,text centered]
\begin{tikzpicture}[node distance = 0,scale = 0.8]
\tikzstyle{every node}=[scale=0.8]
\node(figure-1)[draw=white,scale=0.25] at (0,0){\includegraphics[width=0.62\textwidth]{./Chapter17/Figures/figure-dog-with-hat.png}};
\node(ground-1)[rectangle,rounded corners, minimum width=5cm, minimum height=3.5cm,right of = figure-1, xshift=5cm,fill=blue!15]{};
\node(text-1)[right of = figure-1, xshift=3.6cm,yshift=2cm,scale=1.2]{\textcolor{color_gray}{描述候选池}};
\node(text_1-1)[description, right of = figure-1, xshift=4.2cm,yshift=1.2cm,fill=color_gray!50]{\textcolor{white}{天空中有很多鸟。}};
\node(text_2-1)[description, right of = figure-1, xshift=5.3cm,yshift=0.5cm,fill=color_green]{\textcolor{white}{孩子从河岸上跳下来。}};
\node(text_3-1)[description, right of = figure-1, xshift=4.5cm,yshift=-0.2cm,fill=color_orange]{\textcolor{white}{狗在吐舌头。}};
\node(surd-1)[right of = text_3-1, xshift=2cm,scale=1.5]{\textcolor{red}{$\surd$}};
\node(text_4-1)[description, right of = figure-1, xshift=5.2cm,yshift=-0.9cm,fill=color_blue]{\textcolor{white}{男人戴着眼镜。}};
\node(point-1)[right of = figure-1, xshift=5cm,yshift=-1.4cm,scale=1.5]{...};
\draw[->,very thick](figure-1)to([xshift=-0.1cm]ground-1.west);
\node(figure)[draw=white,scale=0.25]at ([xshift=20.0em]figure-1.east){\includegraphics[width=0.62\textwidth]{./Chapter17/Figures/figure-dog-with-hat.png}};
\node(ground)[rectangle,rounded corners, minimum width=5cm, minimum height=1.5cm,right of = figure, xshift=5cm,yshift=-0.8cm,fill=blue!15]{\large{图片中有\underline{\textcolor{red}{狗}}，\underline{\textcolor{red}{帽子}}，\underline{\quad\ }。}};
\node(dog)[rectangle,rounded corners, minimum width=1cm, minimum height=0.7cm,right of = figure, xshift=3cm,yshift=1.5cm,thick, draw=color_orange,fill=color_orange!50]{狗};
\node(hat)[rectangle,rounded corners, minimum width=1.5cm, minimum height=0.7cm,right of = figure, xshift=4.5cm,yshift=1.5cm,thick, draw=color_green,fill=color_green!50]{帽子};
\draw[->, very thick,color=black!60](figure.east)to([xshift=-0.1cm]dog.west)node[left,xshift=-0.2cm,yshift=-0.1cm,color=black]{图片检测};
\draw[->, very thick,color=black!60]([yshift=-0.1cm]hat.south)to([yshift=0.1cm]ground.north)node[right,xshift=-0.2cm,yshift=0.5cm,color=black]{模板填充};
\end{tikzpicture}
\ No newline at end of file
Chapter17/chapter17.tex
......@@ -311,6 +311,7 @@
%----------------------------------------------------------------------------------------------------
\begin{figure}[htp]
\centering
\input{./Chapter17/Figures/figure-comparison-of-attention-mechanism-of-target-word-bank}
\caption{目标词“bank”注意力机制前后对比}
\label{tab:17-2-3-c}
\end{figure}
......@@ -336,13 +337,14 @@
\parinterval 基于多任务学习的方法通常是把翻译任务与其他视觉任务结合，进行联合训练。在{\chapterfifteen}和{\chaptersixteen}已经提到过多任务学习。一种常见的多任务学习框架是针对多个相关的任务，共享模型的部分参数来学习不同任务之间相似的部分，并通过特定的模块来学习每个任务特有的部分。在多模态机器翻译中，应用多任务学习的主要策略就是将翻译作为主任务，同时设置一些与其他模态相关的子任务，通过这些子任务来辅助源语言理解自身的语言知识。
\parinterval 如图4所示，可以将多模态机器翻译任务分解为两个子任务：机器翻译和图片生成\upcite{DBLP:conf/ijcnlp/ElliottK17}。其中机器翻译作为主任务，图片生成作为子任务，图片生成这里指的是从一个图片描述生成对应图片，对于图片生成任务在后面叙述。通过单个编码器对源语言数据进行建模，然后通过两个解码器（翻译解码器和图像解码器）来学习翻译任务和图像生成任务。顶层任务学习每个任务的独立特征，底层共享参数层能够学习到更丰富的文本特征表示。另外在视觉问答领域有研究表明\upcite{DBLP:conf/nips/LuYBP16}，在多模态任务中，不宜引入多层的注意力，因为多层注意力会导致模型严重的过拟合，从另一角度来说，利用多任务学习的方式，提高模型的泛化能力，也是一种有效防止过拟合现象的方式。类似的思想，也大量使用在多模态自然语言处理中，例如图像描述生成、视觉问答\upcite{DBLP:conf/iccv/AntolALMBZP15}等。
\parinterval 如图\ref{fig:17-13}所示，可以将多模态机器翻译任务分解为两个子任务：机器翻译和图片生成\upcite{DBLP:conf/ijcnlp/ElliottK17}。其中机器翻译作为主任务，图片生成作为子任务，图片生成这里指的是从一个图片描述生成对应图片，对于图片生成任务在后面叙述。通过单个编码器对源语言数据进行建模，然后通过两个解码器（翻译解码器和图像解码器）来学习翻译任务和图像生成任务。顶层任务学习每个任务的独立特征，底层共享参数层能够学习到更丰富的文本特征表示。另外在视觉问答领域有研究表明\upcite{DBLP:conf/nips/LuYBP16}，在多模态任务中，不宜引入多层的注意力，因为多层注意力会导致模型严重的过拟合，从另一角度来说，利用多任务学习的方式，提高模型的泛化能力，也是一种有效防止过拟合现象的方式。类似的思想，也大量使用在多模态自然语言处理中，例如图像描述生成、视觉问答\upcite{DBLP:conf/iccv/AntolALMBZP15}等。
%----------------------------------------------------------------------------------------------------
\begin{figure}[htp]
\centering
\input{./Chapter17/Figures/figure-application-of-multimodal-machine-translation-to-multitask-learning.tex}
\caption{多模态机器翻译多任务学习的应用}
\label{tab:17-2-4-c}
\label{fig:17-13}
\end{figure}
%----------------------------------------------------------------------------------------------------
......@@ -357,12 +359,13 @@
%----------------------------------------------------------------------------------------------------
\begin{figure}[htp]
\centering
\input{./Chapter17/Figures/figure-traditional-methods-of-image-description}
\caption{图像描述传统方法}
\label{tab:17-2-5-c}
\end{figure}
%----------------------------------------------------------------------------------------------------
\parinterval 传统图像描述生成有两种范式：基于检索的方法和基于模板的方法。其中基于检索的方法（图5左）是指在指定的图像描述候选句子中选择其中的句子作为图像的描述，这种方法的弊端是所选择的句子可能会和图像很大程度上不相符。而基于模板的方法（图5右）是指在图像上检测视觉特征，然后把内容填在实现设计好的模板当中，这种方法的缺点是生成的图像描述过于呆板，‘像是在一个模子中刻出来的’说的就是这个意思。近几年来 ，由于卷积神经网络在计算机视觉领域效果显著，而循环神经网络在自然语言处理领域卓有成效，受到机器翻译领域编码器-解码器框架的启发，逐渐的，这种基于卷积神经网络作为编码器编码图像，循环神经网络作为解码器解码描述的编码器-解码器框架成了图像描述任务的基础范式。本章节，从基础的图像描述范式编码器-解码器框架展开\upcite{DBLP:conf/cvpr/VinyalsTBE15,DBLP:conf/icml/XuBKCCSZB15}，从编码器的改进、解码器的改进展开介绍。
\parinterval 传统图像描述生成有两种范式：基于检索的方法和基于模板的方法。其中基于检索的方法（图5左）是指在指定的图像描述候选句子中选择其中的句子作为图像的描述，这种方法的弊端是所选择的句子可能会和图像很大程度上不相符。而基于模板的方法（图5右）是指在图像上检测视觉特征，然后把内容填在实现设计好的模板当中，这种方法的缺点是生成的图像描述过于呆板，“像是在一个模子中刻出来的”说的就是这个意思。近几年来 ，由于卷积神经网络在计算机视觉领域效果显著，而循环神经网络在自然语言处理领域卓有成效，受到机器翻译领域编码器-解码器框架的启发，逐渐的，这种基于卷积神经网络作为编码器编码图像，循环神经网络作为解码器解码描述的编码器-解码器框架成了图像描述任务的基础范式。本章节，从基础的图像描述范式编码器-解码器框架展开\upcite{DBLP:conf/cvpr/VinyalsTBE15,DBLP:conf/icml/XuBKCCSZB15}，从编码器的改进、解码器的改进展开介绍。
%----------------------------------------------------------------------------------------
% NEW SUBSUB-SECTION
......@@ -399,13 +402,14 @@
\parinterval 要想使编码器-解码器框架在图像描述中充分发挥作用，编码器也要更好的表示图像信息。对于编码器的改进，大多也是从这个方向出发。通常，体现在向编码器中添加图像的语义信息\upcite{DBLP:conf/cvpr/YouJWFL16,DBLP:conf/cvpr/ChenZXNSLC17,DBLP:journals/pami/FuJCSZ17}和位置信息\upcite{DBLP:conf/cvpr/ChenZXNSLC17,DBLP:conf/ijcai/LiuSWWY17}。
\parinterval 图像的语义信息一般是指图像中存在的实体、属性、场景等等。如图XX所示，从图像中利用属性或实体检测器提取出“child”、“river”、“bank”等等的属性词和实体词作为图像的语义信息，提取全局的图像特征初始化循环神经网络，再利用注意力机制计算目标词与属性词或实体词之间的注意力权重，根据该权重计算上下文向量，从而将编码语义信息送入解码端\upcite{DBLP:conf/cvpr/YouJWFL16}，在解码‘bank’单词时，会更关注图像语义信息中的‘bank’。当然，除了图像中的实体和属性作为语义信息外，也可以将图片的场景信息也加入到编码器当中\upcite{DBLP:journals/pami/FuJCSZ17}。有关如何做属性、实体和场景的检测，涉及到目标检测任务的工作，例如Faster-RCNN\upcite{DBLP:journals/pami/RenHG017}、YOLO\upcite{DBLP:journals/corr/abs-1804-02767,DBLP:journals/corr/abs-2004-10934}等等,这里不过多赘述。
\parinterval 图像的语义信息一般是指图像中存在的实体、属性、场景等等。如图\ref{fig:17-16}所示，从图像中利用属性或实体检测器提取出“child”、“river”、“bank”等等的属性词和实体词作为图像的语义信息，提取全局的图像特征初始化循环神经网络，再利用注意力机制计算目标词与属性词或实体词之间的注意力权重，根据该权重计算上下文向量，从而将编码语义信息送入解码端\upcite{DBLP:conf/cvpr/YouJWFL16}，在解码‘bank’单词时，会更关注图像语义信息中的‘bank’。当然，除了图像中的实体和属性作为语义信息外，也可以将图片的场景信息也加入到编码器当中\upcite{DBLP:journals/pami/FuJCSZ17}。有关如何做属性、实体和场景的检测，涉及到目标检测任务的工作，例如Faster-RCNN\upcite{DBLP:journals/pami/RenHG017}、YOLO\upcite{DBLP:journals/corr/abs-1804-02767,DBLP:journals/corr/abs-2004-10934}等等,这里不过多赘述。
%----------------------------------------------------------------------------------------------------
\begin{figure}[htp]
\centering
\input{./Chapter17/Figures/figure-the-encoder-explicitly-incorporates-semantic-information}
\caption{编码器“显式”融入语义信息}
\label{tab:17-2-6-c}
\label{fig:17-16}
\end{figure}
%----------------------------------------------------------------------------------------------------
......
bibliography.bib
......@@ -6601,11 +6601,11 @@ author = {Yoshua Bengio and
year = {2012}
}
@article{DBLP:journals/corr/ZhuB17,
@inproceedings{DBLP:journals/corr/ZhuB17,
author = {Jia-Jie Zhu and
Jos{\'{e}} Bento},
title = {Generative Adversarial Active Learning},
journal = {CoRR},
publisher = {CoRR},
volume = {abs/1702.07956},
year = {2017}
}
......@@ -7047,6 +7047,209 @@ author = {Yoshua Bengio and
year = {2018}
}
@inproceedings{Bengio2015ScheduledSF,
title={Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks},
author={Samy Bengio and
Oriol Vinyals and
Navdeep Jaitly and
Noam Shazeer},
publisher = {Conference and Workshop on Neural Information Processing Systems},
pages = {1171--1179},
year = {2015}
}
@inproceedings{Bengio2015ScheduledSF,
title={Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks},
author={Samy Bengio and
Oriol Vinyals and
Navdeep Jaitly and
Noam Shazeer},
publisher = {Conference and Workshop on Neural Information Processing Systems},
pages = {1171--1179},
year = {2015}
}
@inproceedings{Ranzato2016SequenceLT,
title={Sequence Level Training with Recurrent Neural Networks},
author={Marc'Aurelio Ranzato and
Sumit Chopra and
Michael Auli and
Wojciech Zaremba},
publisher={International Conference on Learning Representations},
year={2016}
}
@inproceedings{DBLP:conf/nips/GoodfellowPMXWOCB14,
author = {Ian J. Goodfellow and
Jean Pouget-Abadie and
Mehdi Mirza and
Bing Xu and
David Warde-Farley and
Sherjil Ozair and
Aaron C. Courville and
Yoshua Bengio},
title = {Generative Adversarial Nets},
publisher = {Conference on Neural Information Processing Systems},
pages = {2672--2680},
year = {2014}
}
@inproceedings{DBLP:conf/acl/ShenCHHWSL16,
author = {Shiqi Shen and
Yong Cheng and
Zhongjun He and
Wei He and
Hua Wu and
Maosong Sun and
Yang Liu},
title = {Minimum Risk Training for Neural Machine Translation},
publisher = {Annual Meeting of the Association for Computational Linguistics},
year = {2016},
}
@inproceedings{DBLP:conf/acl/PapineniRWZ02,
author = {Kishore Papineni and
Salim Roukos and
Todd Ward and
Wei-jing Zhu},
title = {Bleu: a Method for Automatic Evaluation of Machine Translation},
pages = {311--318},
publisher = {Annual Meeting of the Association for Computational Linguistics},
year = {2002}
}
@inproceedings{doddington2002automatic,
title={Automatic evaluation of machine translation quality using n-gram co-occurrence statistics},
publisher={Proceedings of the second international conference on Human Language Technology Research},
author={Doddington, George},
pages={138--145},
year={2002}
}
@inproceedings{snover2006study,
title={A study of translation edit rate with targeted human annotation},
author={Snover, Matthew and Dorr, Bonnie and Schwartz, Richard and Micciulla, Linnea and Makhoul, John},
publisher={Proceedings of association for machine translation in the Americas},
volume={200},
number={6},
year={2006}
}
@inproceedings{lavie2009meteor,
title={The METEOR metric for automatic evaluation of machine translation},
author={Lavie, Alon and Denkowski, Michael J},
publisher={Machine translation},
volume={23},
number={2-3},
pages={105--115},
year={2009}
}
@inproceedings{bahdanau2014neural,
author = {Dzmitry Bahdanau and
Kyunghyun Cho and
Yoshua Bengio},
title = {Neural Machine Translation by Jointly Learning to Align and Translate},
publisher = {International Conference on Learning Representations},
year = {2015}
}
@inproceedings{koehn2003statistical,
author = {Philipp Koehn and
Franz Josef Och and
Daniel Marcu},
title = {Statistical Phrase-Based Translation},
publisher = {Annual Meeting of the Association for Computational Linguistics},
year = {2003}
}
@inproceedings{smith2006minimum,
author = {David A. Smith and
Jason Eisner},
title = {Minimum Risk Annealing for Training Log-Linear Models},
publisher = {Annual Meeting of the Association for Computational Linguistics},
year = {2006}
}
@inproceedings{he2012maximum,
title={Maximum expected bleu training of phrase and lexicon translation models},
author={He, Xiaodong and Deng, Li},
publisher={Annual Meeting of the Association for Computational Linguistics},
pages={292--301},
year={2012}
}
@inproceedings{DBLP:conf/acl/GaoHYD14,
author = {Jianfeng Gao and
Xiaodong He and
Wen{-}tau Yih and
Li Deng},
title = {Learning Continuous Phrase Representations for Translation Modeling},
pages = {699--709},
publisher = {Annual Meeting of the Association for Computational Linguistics},
year = {2014}
}
@book{sutton2018reinforcement,
title={Reinforcement learning: An introduction},
author={Richard S. Sutton and Andrew G. Barto},
year={2018},
publisher={MIT press}
}
@inproceedings{DBLP:journals/nature/SilverHMGSDSAPL16,
author = {David Silver and
Aja Huang and
Chris J. Maddison and
Arthur Guez and
Laurent Sifre and
George van den Driessche and
Julian Schrittwieser and
Ioannis Antonoglou and
Vedavyas Panneershelvam and
Marc Lanctot and
Sander Dieleman and
Dominik Grewe and
John Nham and
Nal Kalchbrenner and
Ilya Sutskever and
Timothy P. Lillicrap and
Madeleine Leach and
Koray Kavukcuoglu and
Thore Graepel and
Demis Hassabis},
title = {Mastering the game of Go with deep neural networks and tree search},
publisher = {Nature},
volume = {529},
number = {7587},
pages = {484--489},
year = {2016}
}
@inproceedings{DBLP:conf/icml/ZarembaMJF16,
author = {Wojciech Zaremba and
Tomas Mikolov and
Armand Joulin and
Rob Fergus},
title = {Learning Simple Algorithms from Examples},
series = {{JMLR} Workshop and Conference Proceedings},
volume = {48},
pages = {421--429},
publisher = {International Conference on Machine Learning},
year = {2016}
}
@inproceedings{DBLP:conf/icml/NgHR99,
author = {Andrew Y. Ng and
Daishi Harada and
Stuart J. Russell},
title = {Policy Invariance Under Reward Transformations: Theory and Application
to Reward Shaping},
pages = {278--287},
publisher = {International Conference on Machine Learning},
year = {1999}
}
%%%%% chapter 13------------------------------------------------------
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
......
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