Vectorized implementation
Vectorization
Vectorization refers to a powerful way to speed up your algorithms. Numerical computing and parallel computing researchers have put decades of work into making certain numerical operations (such as matrix-matrix multiplication, matrix-matrix addition, matrix-vector multiplication) fast. The idea of vectorization is that we would like to express our learning algorithms in terms of these highly optimized operations.
More generally, a good rule-of-thumb for coding Matlab/Octave is:
- Whenever possible, avoid using explicit for-loops in your code.
A large part of vectorizing our Matlab/Octave code will focus on getting rid of for loops, since this lets Matlab/Octave extract more parallelism from your code, while also incurring less computational overhead from the interpreter.
多用向量运算,别把向量拆成标量然后再循环
Logistic Regression Vectorization Example
Consider training a logistic regression model using batch gradient ascent. Suppose our hypothesis is
where we let
, so that
and
, and
is our intercept term. We have a training set
of
examples, and the batch gradient ascent update rule is
, where
is the log likelihood and
is its derivative.
We thus need to compute the gradient:
Further, suppose the Matlab/Octave variable y is a row vector of the labels in the training set, so that the variable y(i) is
.
Here's truly horrible, extremely slow, implementation of the gradient computation:
% Implementation
grad = zeros(n+,);
for i=:m,
h = sigmoid(theta'*x(:,i));
temp = y(i) - h;
for j=:n+,
grad(j) = grad(j) + temp * x(j,i);
end;
end;The two nested for-loops makes this very slow. Here's a more typical implementation, that partially vectorizes the algorithm and gets better performance:
% Implementation
grad = zeros(n+,);
for i=:m,
grad = grad + (y(i) - sigmoid(theta'*x(:,i)))* x(:,i);
end;
Neural Network Vectorization
Forward propagation
Consider a 3 layer neural network (with one input, one hidden, and one output layer), and suppose x is a column vector containing a single training example
. Then the forward propagation step is given by:
This is a fairly efficient implementation for a single example. If we have m examples, then we would wrap a for loop around this.
% Unvectorized implementation
for i=:m,
z2 = W1 * x(:,i) + b1;
a2 = f(z2);
z3 = W2 * a2 + b2;
h(:,i) = f(z3);
end;For many algorithms, we will represent intermediate stages of computation via vectors. For example, z2, a2, and z3 here are all column vectors that're used to compute the activations of the hidden and output layers. In order to take better advantage of parallelism and efficient matrix operations, we would like to have our algorithm operate simultaneously on many training examples. Let us temporarily ignore b1 and b2 (say, set them to zero for now). We can then implement the following:
% Vectorized implementation (ignoring b1, b2)
z2 = W1 * x;
a2 = f(z2);
z3 = W2 * a2;
h = f(z3)In this implementation, z2, a2, and z3 are all matrices, with one column per training example.
A common design pattern in vectorizing across training examples is that whereas previously we had a column vector (such as z2) per training example, we can often instead try to compute a matrix so that all of these column vectors are stacked together to form a matrix. Concretely, in this example, a2 becomes a s2 by m matrix (where s2 is the number of units in layer 2 of the network, and m is the number of training examples). And, the i-th column of a2 contains the activations of the hidden units (layer 2 of the network) when the i-th training example x(:,i) is input to the network.
% Inefficient, unvectorized implementation of the activation function
function output = unvectorized_f(z)
output = zeros(size(z))
for i=:size(z,),
for j=:size(z,),
output(i,j) = /(+exp(-z(i,j)));
end;
end;
end % Efficient, vectorized implementation of the activation function
function output = vectorized_f(z)
output = ./(+exp(-z)); % "./" is Matlab/Octave's element-wise division operator.
endFinally, our vectorized implementation of forward propagation above had ignored b1 and b2. To incorporate those back in, we will use Matlab/Octave's built-in repmat function. We have:
% Vectorized implementation of forward propagation
z2 = W1 * x + repmat(b1,,m);
a2 = f(z2);
z3 = W2 * a2 + repmat(b2,,m);
h = f(z3)repmat !!矩阵变形!!
Backpropagation
We are in a supervised learning setting, so that we have a training set
of m training examples. (For the autoencoder, we simply set y(i) = x(i), but our derivation here will consider this more general setting.)
we had that for a single training example (x,y), we can compute the derivatives as
Here,
denotes element-wise product. For simplicity, our description here will ignore the derivatives with respect to b(l), though your implementation of backpropagation will have to compute those derivatives too.
gradW1 = zeros(size(W1));
gradW2 = zeros(size(W2));
for i=:m,
delta3 = -(y(:,i) - h(:,i)) .* fprime(z3(:,i));
delta2 = W2'*delta3(:,i) .* fprime(z2(:,i)); gradW2 = gradW2 + delta3*a2(:,i)';
gradW1 = gradW1 + delta2*a1(:,i)';
end;This implementation has a for loop. We would like to come up with an implementation that simultaneously performs backpropagation on all the examples, and eliminates this for loop.
To do so, we will replace the vectors delta3 and delta2 with matrices, where one column of each matrix corresponds to each training example. We will also implement a function fprime(z) that takes as input a matrix z, and applies
element-wise.
Sparse autoencoder
When performing backpropagation on a single training example, we had taken into the account the sparsity penalty by computing the following:
也就是不要用循环一个样本一个样本的去更新参数,而是要将样本组织成矩阵的形式,应用矩阵运算,提高效率。
Vectorized implementation的更多相关文章
- DL三(向量化编程 Vectorized implementation)
向量化编程实现 Vectorized implementation 一向量化编程 Vectorization 1.1 基本术语 向量化 vectorization 1.2 向量化编程(Vectoriz ...
- 机器学习公开课笔记(4):神经网络(Neural Network)——表示
动机(Motivation) 对于非线性分类问题,如果用多元线性回归进行分类,需要构造许多高次项,导致特征特多学习参数过多,从而复杂度太高. 神经网络(Neural Network) 一个简单的神经网 ...
- 转载 Deep learning:一(基础知识_1)
前言: 最近打算稍微系统的学习下deep learing的一些理论知识,打算采用Andrew Ng的网页教程UFLDL Tutorial,据说这个教程写得浅显易懂,也不太长.不过在这这之前还是复习下m ...
- Deep learning:一(基础知识_1)
本文纯转载: 主要是想系统的跟tornadomeet的顺序走一遍deeplearning; 前言: 最近打算稍微系统的学习下deep learing的一些理论知识,打算采用Andrew Ng的网页教程 ...
- machine learning 之 Neural Network 1
整理自Andrew Ng的machine learning课程week 4. 目录: 为什么要用神经网络 神经网络的模型表示 1 神经网络的模型表示 2 实例1 实例2 多分类问题 1.为什么要用神经 ...
- Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Regularization)
声明:所有内容来自coursera,作为个人学习笔记记录在这里. Regularization Welcome to the second assignment of this week. Deep ...
- Coursera机器学习+deeplearning.ai+斯坦福CS231n
日志 20170410 Coursera机器学习 2017.11.28 update deeplearning 台大的机器学习课程:台湾大学林轩田和李宏毅机器学习课程 Coursera机器学习 Wee ...
- Neural Networks and Deep Learning 课程笔记(第三周)浅层神经网络(Shallow neural networks)
3.1 神经网络概述(Neural Network Overview ) (神经网络中,我们要反复计算a和z,最终得到最后的loss function) 3.2 神经网络的表示(Neural Netw ...
- [UFLDL] Basic Concept
博客内容取材于:http://www.cnblogs.com/tornadomeet/archive/2012/06/24/2560261.html 参考资料: UFLDL wiki UFLDL St ...
随机推荐
- Objective-C 小记(10)__weak
本文使用的 runtime 版本为 objc4-706. __weak 修饰的指针最重要的特性是其指向的对象销毁后,会自动置为 nil,这个特性的实现完全是依靠运行时的.实现思路是非常简单的,对于下面 ...
- WLAN 基础架构功能
WLAN 基础架构功能 Android WLAN 框架可帮助用户连接到优质 WLAN 网络(在有可用 WLAN 网络且需要连接到这类网络的情况下).Android 可通过多种方式来实现这一点: 打开网 ...
- BZOJ 3674: 可持久化并查集模板
Code: #include <cstdio> #include <algorithm> #include <cstring> #include <strin ...
- Mysql 主从主主复制总结(详细)
环境:Centos 6.9,Mysql 8.0 配置准备:1.为Centos配置好网络,使用远程工具连接. 2.安装mysql,途径不限.mysql8.0和5.7区别不大,8.0在配置主从的时候默认开 ...
- websocket调试工具
http://www.blue-zero.com/WebSocket/ wss://yy.xxx.com/video/websocket/client.ws
- 使用 init-runonce脚本创建一个 openstack云项目
source /etc/kolla/admin-openrc.sh cd /usr/share/kolla-ansible ./init-runonce 报错内容 Traceback (most re ...
- Linux LVM在线扩容
环境: 虚拟化环境,SUSE Linux Enterprise Server 11sp3,直接把虚拟磁盘从100G改成150G. 现有的LVM是100G,/home 的LV需要再加50G. 步骤: f ...
- 洛谷—— P1003 铺地毯
https://www.luogu.org/problem/show?pid=1003 题目描述 为了准备一个独特的颁奖典礼,组织者在会场的一片矩形区域(可看做是平面直角坐标系的第一象限)铺上一些矩形 ...
- gluPerspective和gluLookAt的关系
参考文章 GL学习笔记(2) - 终于搞明白gluPerspective和gluLookAt的关系了(zz) gluPerspective的具体含义 解密--神秘的gluPerspective 函数原 ...
- ArcGIS api for javascript——加入地图并显示当前地图范围
描述 这个示例使用Map.extent property属性接收地图范围的左下角和右上角坐标 "书签". 使用下列行创建地图: var map = new esri.Map(&qu ...

, so that
and
, and
is our intercept term. We have a training set
of
examples, and the batch gradient ascent update rule is
, where
is the log likelihood and
is its derivative.
.
. Then the forward propagation step is given by:
of m training examples. (For the autoencoder, we simply set y(i) = x(i), but our derivation here will consider this more general setting.)
denotes element-wise product. For simplicity, our description here will ignore the derivatives with respect to b(l), though your implementation of backpropagation will have to compute those derivatives too.
element-wise.