exacity / deeplearningbook-chinese

Deep Learning Book Chinese Translation
https://exacity.github.io/deeplearningbook-chinese
35.62k stars 9.12k forks source link

Chapter1 #2

Closed SwordYork closed 7 years ago

SwordYork commented 7 years ago

关于第一章的各种建议,可以在这个issue下提

huangpingchun commented 7 years ago

1.2.1小节 (257行) “最近才成为所谓的``深度学习''。 这个邻域已经更换了很多名称,反映了不同的研究人员和不同观点的影响。” 还有一个 “邻域”的错字。

SwordYork commented 7 years ago

惭愧,以后不会有了,我grep过了。 刚刚发现Chapter7/regularization.tex:440 也有这种错误。

xiaomingabc commented 7 years ago

提一个小问题,第一章 前言 第二段的第二行 “这距造出一台还有一百年多年之久” 好像是个病句

SwordYork commented 7 years ago

@xiaomingabc 确实是个病句,修改语言时会一同修改掉。多谢!

bengordai commented 7 years ago

第9段(第3页) A simple machine learning algorithm called naive Bayes can separate legitimate e-mail from spam e-mail 叫做朴素贝叶斯(naive Bayes) 的简单机器学习算法可以从垃圾邮件中区分合法的电子邮件。

我理解为tell good from bad,是在普通邮件中区分出合法邮件与垃圾邮件这样的意思。

SwordYork commented 7 years ago

@bengordai 很有道理!我中文思维了,我是从一堆垃圾邮件中区分合法邮件的感觉。多谢!你不说我还真不会改出来。

linzhp commented 7 years ago

"The true challenge to artificial intelligence proved to be solving the tasks that are easy for people to perform but hard for people to describe formally—problems that we solve intuitively, that feel automatic, like recognizing spoken words or faces in images"

这句话是不是可以考虑这样翻译顺一些?

人工智能的真正挑战其实是解决那些难以形式化描述的任务,比如识别人们说的话或图像中的脸。这些我们能通过直觉轻而易举地解决。

prove有turn out的意思,不一定要翻译成证明 automatic有不由自主的意思,不一定要翻译成自动

xiaomingabc commented 7 years ago

agree

------------------ 原始邮件 ------------------ 发件人: "Zhongpeng Lin";notifications@github.com; 发送时间: 2017年1月9日(星期一) 上午9:15 收件人: "exacity/deeplearningbook-chinese"deeplearningbook-chinese@noreply.github.com; 抄送: "xu jia liang"2252128507@qq.com; "Mention"mention@noreply.github.com; 主题: Re: [exacity/deeplearningbook-chinese] Chapter1 (#2)

"The true challenge to artificial intelligence proved to be solving the tasks that are easy for people to perform but hard for people to describe formally—problems that we solve intuitively, that feel automatic, like recognizing spoken words or faces in images"

这句话是不是可以考虑这样翻译顺一些?

人工智能的真正挑战其实是解决那些难以形式化描述的任务,比如识别人们说的话或图像中的脸。这些我们能通过直觉轻而易举地解决。

prove有turn out的意思,不一定要翻译成证明 automatic有不由自主的意思,不一定要翻译成自动

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

linzhp commented 7 years ago

the set of chess pieces and allowable moves可以翻译成“棋子及其走法”

SwordYork commented 7 years ago

@linzhp 收到,多谢!

linzhp commented 7 years ago

An auto encoder is the combination of an encoder function that converts the input data into a different representation, and a decoder function that converts the new representation back into the original format. 翻译中少了个“的”:自动编码器组合了将输入转换到不同表示编码器函数和将新的表示转回原来形式的解码器函数。

或者可以考虑翻译成:

自动编码器结合了编码器函数来将输入转换到不同表示,以及解码器函数来将新的表示转回原来的形式。

linzhp commented 7 years ago

They may also exist as constructs in the human mind that provide useful simplifying explanations or inferred causes of the observed data 应该翻译成: 它们还可能以概念的形式存在于人们的头脑中,对观测到的数据提供简洁有用的解释或推断其原因。

Construct有概念的意思

SwordYork commented 7 years ago

@linzhp 您好,多谢您的建议!这些建议都很有用。如果您有时间,是否可以在这个网页批注。 这样更方便一点。这是第一章的,其他章只要把后面的地址换掉就行。我们下一阶段将采用这种形式,目前还不稳定,所以尚未公布这种批注方式,如果您有时间,可以先尝试一下。

linzhp commented 7 years ago

好的,我试试

SwordYork commented 7 years ago

@linzhp 非常感谢细心的批注,我们在一段时间后会综合意见一起合并。由于您已经对第一章做了大量批注,所以我已经列入第一章中英对比的校对者。不知道您是否有兴趣进行其他章节的校对?

linzhp commented 7 years ago

没问题。这本书我也没读过,正好一边读一边做点贡献

linzhp commented 7 years ago

有点好奇:这本书的英文网站上写着说这书的中文版版权已经被邮电出版社买了。你们这个翻译会不会有麻烦?跟邮电出版社沟通过了吗?

liber145 commented 7 years ago

@linzhp 😄 不会有麻烦的。我们得到了邮电出版社的授权。

taiyun commented 7 years ago

@linzhp 翻译最终纸质版会在人民邮电出版社出版~

zdx3578 commented 7 years ago

knowledge graph 知识图谱 ; knowledge base:知识为基础 ; 有朋友建议knowledge base 翻译为知识库

SwordYork commented 7 years ago

@zdx3578 多谢,因为书中只有一处出现knowledge graph,所以忘记区分了。我已经把knowledge base 翻译为知识库了,多谢!

germany-zhu commented 7 years ago

英文版第三页第三段,我作为语音方向的研究生校正一下语音术语 For example, a useful feature for speaker identification from sound is the pitch. The pitch can be formally specified—it is the lowest frequency major peak of the spectrogram. It is useful for speaker identification because it is determined by the size of the vocal tract, and therefore gives a strong clue as to whether the speaker is a man, woman, or child.例如,对基于声音的说话人确认来说,一个有用特征是基音。基音是有其正式定义的:它是语谱图上频率最低的主峰。基音对说话人确认很重要,因为它是由声道大小决定的,因而为判断说话人是男人、女人还是小孩提供了很强的线索。

SwordYork commented 7 years ago

@germany-zhu 非常感谢!果然专业!

SwordYork commented 7 years ago

@germany-zhu 不好意思,我没找到原文,最新的这一段已经改了。你是在哪看到的?

germany-zhu commented 7 years ago

@SwordYork html英文原文在第三页倒数第二段,中文翻译在图1.1下边(中文pdf的第三页)。我是在这里看的最新版本翻译,之前有人修改过,不过还是不对,把pitch的整句扔了,变成“一个有用的特征是对其声道大小的估计”。公共修改的页面我打开太卡了,所以才在这里回复,结果搞到沟通原文好不方便。。。

SwordYork commented 7 years ago

@germany-zhu 没事,第一章公共修改页面批注太多导致很卡。 你是在这看的吗?这里没有那句话。我觉得你的原文PDF不是最新的(不是翻译的PDF)。 http://www.deeplearningbook.org/contents/intro.html 原文如下,

For example, a useful feature for speaker identification from sound is an estimate of the size of speaker’s vocal tract. It therefore gives a strong clue as to whether the speaker is a man, woman, or child.

germany-zhu commented 7 years ago

@SwordYork 我的原文还真的不是最新,你们是怎么读的呢?网页那里字太小了。 那就只需要把开头改为“对声音的说话人识别来说”。上次我说是说话人确认,不对。 虽然我是读到这中文(特征是声道大小的估计)觉得别扭才找的原文,不知你们觉得不,但是最新的原文就是这么写的,除了直译我也想不出更好的译法了。其实原文不想涉及语音具体知识,把pitch定义删了就可以:a useful feature for speaker identification from sound is the pitch,which is determined by the size of the vocal tract。 好吧,不纠结了,也不是书的重点。希望之后能给出好的修改建议吧。

SwordYork commented 7 years ago

@germany-zhu 你可以下载这个,或者浏览器放大网页,或者直接网页打印为pdf。 https://github.com/HFTrader/DeepLearningBook

把开头改为“对声音的说话人识别来说” 感觉不通顺了,这里我们先按原来的来,之后再看看。 多谢建议,我也觉得作者是为了避免太专业的术语,把原来的写的通俗一点了。

germany-zhu commented 7 years ago

@SwordYork 恩恩不改也好,“对声音的说话人识别来说” 不学语音的人看不习惯

germany-zhu commented 7 years ago

原文第8页首段结尾:

In this case, the graph of concepts only includes two layers—a layer for eyes and a layer for faces—but the graph of computations includes 2n layers if we refine our estimate of each concept given the other n times.

现在的翻译:

但如果我们根据每个概念给出的其他 n 次估计进行细化,计算的图将包括 2n 层。

原文意思应该是:但如果我们提炼每个概念的估计需要额外的n次计算,计算的图将包含2n层。 具体译法你们再润色吧。

germany-zhu commented 7 years ago

第11页第2段:

Part II describes the most established deep learning algorithms that are essentially solved technologies.

原文是定语从句,翻译为:第二部分介绍最成熟的深度学习算法,这些技术基本上已经得到解决。 现在的翻译是并列两个短语:“技术和算法”。 翻译很好,我就挑点小毛病。

germany-zhu commented 7 years ago

It appeals to a more general principle of learning multiple levels of composition, which can be applied in machine learning frameworks that are not necessarily neurally inspired.

深度学习启发于学习多层组合的更普遍的原理,可以应用到那些并非受神经科学启发的机器学习框架中。 有些意思翻译错了。

germany-zhu commented 7 years ago

中文第14页,整流线性单元,真的没那样说的,我常见是修正线性单元。 ReLU的翻译,摘自百度百科:线性整流函数(Rectified Linear Unit, ReLU),又称修正线性单元, 是一种人工神经网络中常用的激活函数。

SwordYork commented 7 years ago

@germany-zhu 收到,关于ReLU我们已经分析过,不管外界怎么说,我们坚持使用整流线性单元,因为它表示 整流后的线性单元,而不是线性的整流函数。我们认为别人翻译有问题,而修正不太好,你懂的。

germany-zhu commented 7 years ago

@SwordYork 有想法的译者们!

SwordYork commented 7 years ago

@germany-zhu 非常感谢!你提出的3处地方确实都存在问题,我修改了一下。尽管多提意见,我们不会觉得麻烦。多点意见,我们也好判断。

SwordYork commented 7 years ago

我认为第一章大问题没有了,但存在的问题很难改进了。 大家可以先不管这一章了,作为一个节点,我先关了这个issue,之后重开新的。

psychelzh commented 6 years ago

感谢 @SwordYork 的出色作品,质量非常好,受益匪浅啊。 但有一个小小的问题,representation一般都是翻译成“表征”的,本书中翻译成“表示”,有什么道理在里面吗?

SwordYork commented 6 years ago

@psychelzh 你好,关于这个问题,我们在第一个issue讨论过,感兴趣的话可以看一下,谢谢你的意见。

psychelzh commented 6 years ago

@SwordYork 谢谢提醒,我知道了。原来表示这个翻译更常见啊。