-
Notifications
You must be signed in to change notification settings - Fork 339
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
5 changed files
with
11 additions
and
12 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
|
@@ -53,8 +53,7 @@ \chapter{深度学习} | |
深度神经网络背后核心的原理,并将这些原理用在一个 MNIST 问题的解决中,方便我们的 | ||
理解。换句话说,本章目标不是将最前沿的神经网络展示给你看。包括前面的章节,我们都 | ||
是聚焦在基础上,这样读者就能够做好充分的准备来掌握众多的不断涌现的深度学习领域最 | ||
新工作。本章仍然在Beta版。期望读者指出笔误,bug,小错和主要的误解。如果你发现了 | ||
可疑的地方,请直接联系 [email protected]。 | ||
新工作。 | ||
|
||
\section{介绍卷积网络} | ||
\label{sec:convolutional_networks} | ||
|
@@ -776,13 +775,13 @@ \subsection*{问题} | |
学习变得非常容易!\\ | ||
|
||
\textbf{这些网络有多深?} 把卷积--\gls*{pooling}层算作一个层,我们最终的架构有 $4$ 个\gls*{hidden-layer}。 | ||
这样的一个真的应该被称为一个深度网络吗?当然,$4$ 个\gls*{hidden-layer}远远多于我们前面学习的 | ||
这样的一个网络真的应该被称为一个深度网络吗?当然,$4$ 个\gls*{hidden-layer}远远多于我们前面学习的 | ||
浅层网络。那些网络大部分只有一个\gls*{hidden-layer},或者偶尔有 $2$ 个\gls*{hidden-layer}。另一方面,2015年 | ||
使用最先进技术的深度网络有时候有几十个\gls*{hidden-layer}。我偶尔听到有人采取“更比你更深”的 | ||
态度,认为如果你没有跟上在隐层数目方面的攀比,那么你真的没有在做深度学习。我不赞 | ||
同这样的态度,部分因为它使得深度学习的定义像是时刻就有结果的事。深度学习中实际的 | ||
突破是认识到它超过浅的 $1$、$2$ 层的网络是切实可行的,这样的浅层网络直到 2000 年 | ||
中都占据优势。这确实是一个重大的突破,开启了更多有特殊意义的模型的探索。但除这之 | ||
突破是认识到它超过浅的 $1$、$2$ 层的网络是切实可行的,这样的浅层网络直到 00 年代 | ||
中期都占据优势。这确实是一个重大的突破,开启了更多有特殊意义的模型的探索。但除这之 | ||
外,层的数目并不是主要的基本利益关系。更确切地说,使用更深层的网络是一种用来帮助 | ||
实现其他目标工具~——~例如更好的分类精确率。 | ||
\\ | ||
|
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters