Overview

In the previous sections, you constructed a 3-layer neural network comprising an input, hidden and output layer. While fairly effective for MNIST, this 3-layer model is a fairly shallow network; by this, we mean that the features (hidden layer activations a(2)) are computed using only "one layer" of computation (the hidden layer).

In this section, we begin to discuss deep neural networks, meaning ones in which we have multiple hidden layers; this will allow us to compute much more complex features of the input. Because each hidden layer computes a non-linear transformation of the previous layer, a deep network can have significantly greater representational power (i.e., can learn significantly more complex functions) than a shallow one.

Note that when training a deep network, it is important to use a non-linear activation function in each hidden layer. This is because multiple layers of linear functions would itself compute only a linear function of the input (i.e., composing multiple linear functions together results in just another linear function), and thus be no more expressive than using just a single layer of hidden units.

Advantages of deep networks

The primary advantage is that it can compactly represent a significantly larger set of fuctions than shallow networks. Formally, one can show that there are functions which a k-layer network can represent compactly (with a number of hidden units that is polynomial in the number of inputs), that a (k − 1)-layer network cannot represent unless it has an exponentially large number of hidden units.

By using a deep network, in the case of images, one can also start to learn part-whole decompositions. For example, the first layer might learn to group together pixels in an image in order to detect edges (as seen in the earlier exercises). The second layer might then group together edges to detect longer contours, or perhaps detect simple "parts of objects." An even deeper layer might then group together these contours or detect even more complex features.

Finally, cortical computations (in the brain) also have multiple layers of processing. For example, visual images are processed in multiple stages by the brain, by cortical area "V1", followed by cortical area "V2" (a different part of the brain), and so on.

Difficulty of training deep architectures

The main learning algorithm that researchers were using was to randomly initialize the weights of a deep network, and then train it using a labeled training set using a supervised learning objective, for example by applying gradient descent to try to drive down the training error. However, this usually did not work well. There were several reasons for this.

Availability of data
Local optima
Diffusion of gradients

when using backpropagation to compute the derivatives, the gradients that are propagated backwards (from the output layer to the earlier layers of the network) rapidly diminish in magnitude as the depth of the network increases. As a result, the derivative of the overall cost with respect to the weights in the earlier layers is very small.(深度神经网络的前几层) Thus, when using gradient descent, the weights of the earlier layers change slowly, and the earlier layers fail to learn much. This problem is often called the "diffusion of gradients."

A closely related problem to the diffusion of gradients is that if the last few layers in a neural network have a large enough number of neurons, it may be possible for them to model the labeled data alone without the help of the earlier layers. Hence, training the entire network at once with all the layers randomly initialized ends up giving similar performance to training a shallow network (the last few layers) on corrupted input (the result of the processing done by the earlier layers).

Greedy layer-wise training

the main idea is to train the layers of the network one at a time, so that we first train a network with 1 hidden layer, and only after that is done, train a network with 2 hidden layers, and so on. At each step, we take the old network with k − 1 hidden layers, and add an additional k-th hidden layer (that takes as input the previous hidden layer k − 1 that we had just trained). Training can either be supervised (say, with classification error as the objective function on each step), but more frequently it is unsupervised (as in an autoencoder; details to provided later). The weights from training the layers individually are then used to initialize the weights in the final/overall deep network, and only then is the entire architecture "fine-tuned" (i.e., trained together to optimize the labeled training set error).

Availability of data

While labeled data can be expensive to obtain, unlabeled data is cheap and plentiful. The promise of self-taught learning is that by exploiting the massive amount of unlabeled data, we can learn much better models. By using unlabeled data to learn a good initial value for the weights in all the layers (except for the final classification layer that maps to the outputs/predictions), our algorithm is able to learn and discover patterns from massively more amounts of data than purely supervised approaches. This often results in much better classifiers being learned.

Better local optima

After having trained the network on the unlabeled data, the weights are now starting at a better location in parameter space than if they had been randomly initialized. We can then further fine-tune the weights starting from this location. Empirically, it turns out that gradient descent from this location is much more likely to lead to a good local minimum, because the unlabeled data has already provided a significant amount of "prior" information about what patterns there are in the input data.

Deep Networks : Overview的更多相关文章

  1. Deep Learning 8_深度学习UFLDL教程:Stacked Autocoders and Implement deep networks for digit classification_Exercise(斯坦福大学深度学习教程)

    前言 1.理论知识:UFLDL教程.Deep learning:十六(deep networks) 2.实验环境:win7, matlab2015b,16G内存,2T硬盘 3.实验内容:Exercis ...

  2. Initialization of deep networks

    Initialization of deep networks 24 Feb 2015Gustav Larsson As we all know, the solution to a non-conv ...

  3. 基于pytorch实现HighWay Networks之Train Deep Networks

    (一)Highway Networks 与 Deep Networks 的关系 理论实践表明神经网络的深度是至关重要的,深层神经网络在很多方面都已经取得了很好的效果,例如,在1000-class Im ...

  4. 论文笔记:SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks

    SiamRPN++: Evolution of Siamese Visual Tracking with Very Deep Networks 2019-04-02 12:44:36 Paper:ht ...

  5. 论文笔记:Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

    Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks ICML 2017 Paper:https://arxiv.org/ ...

  6. 【DeepLearning】Exercise: Implement deep networks for digit classification

    Exercise: Implement deep networks for digit classification 习题链接:Exercise: Implement deep networks fo ...

  7. 深度学习材料:从感知机到深度网络A Deep Learning Tutorial: From Perceptrons to Deep Networks

    In recent years, there’s been a resurgence in the field of Artificial Intelligence. It’s spread beyo ...

  8. Deep Networks for Image Super-Resolution with Sparse Prior

    深度学习中潜藏的稀疏表达 Deep Networks for Image Super-Resolution with Sparse Prior http://www.ifp.illinois.edu/ ...

  9. Training Very Deep Networks

    Rupesh Kumar SrivastavaKlaus Greff ̈J urgenSchmidhuberThe Swiss AI Lab IDSIA / USI / SUPSI{rupesh, k ...

随机推荐

  1. mysql读写分离的解决方案

    来源于网上整理 http://yanwt.iteye.com/blog/1460780 现有三种解决方式实现mysql读写分离 1 程序修改mysql操作类 优点:直接和数据库通信,简单快捷的读写分离 ...

  2. codeforces 501 B Misha and Changing Handles 【map】

    题意:给出n个名字变化,问一个名字最后变成了什么名字 先用map顺着做的,后来不对, 发现别人是将变化后的那个名字当成键值来做的,最后输出的时候先输出second,再输出first 写一下样例就好理解 ...

  3. tf.nn.top_k(input, k, name=None)和tf.nn.in_top_k(predictions, targets, k, name=None)

    tf.nn.top_k(input, k, name=None) 这个函数的作用是返回 input 中每行最大的 k 个数,并且返回它们所在位置的索引. input: 一个张量,数据类型必须是以下之一 ...

  4. 创建vue.js项目

    vue init webpack test cd test npm install 或者 cnpm install test npm run dev

  5. [USACO07MAR]每月的费用Monthly Expense

    题目:POJ3273.洛谷P2884. 题目大意:有n个数,要分成m份,每份的和要尽可能小,求这个情况下和最大的一份的和. 解题思路:二分答案,对每个答案进行贪心判断,如果最后得出份数>m,则说 ...

  6. Django初学习程序大致流程

  7. subline Text3 插件安装

    --没有解决,换了vscode 安装Package Control 这是必须的步骤,安装任何插件之前需要安装这个 自动安装的方法最方便,只需要在控制台(不是win的控制台,而是subline 的)里粘 ...

  8. caioj 1087 动态规划入门(非常规DP11:潜水员)(二维背包)

    这道题的难点在于价值可以多. 这道题我一开始用的是前面的状态推现在的状态 实现比较麻烦,因为价值可以多,所以就设最大价值 为题目给的最大价值乘以10 #include<cstdio> #i ...

  9. Java基础学习总结(18)——网络编程

    一.网络基础概念 首先理清一个概念:网络编程 != 网站编程,网络编程现在一般称为TCP/IP编程. 二.网络通信协议及接口 三.通信协议分层思想 四.参考模型 五.IP协议 每个人的电脑都有一个独一 ...

  10. 进程:linux用户态-内核态

    用户态:Ring3运行于用户态的代码则要受到处理器的诸多检查,它们只能访问映射其地址空间的页表项中规定的在用户态下可访问页面的虚拟地址,且只能对任务状态段(TSS)中I/O许可位图(I/O Permi ...