Engineering and Information Security Stuff

Search This Blog

Safe JavaScript Templating Portal

on October 01, 2015

Get link
Facebook
X
Pinterest
Email
Other Apps

Finally, we have launched the portal of Safe JavaScript Templating, and you can try the demo here!

Get link
Facebook
X
Pinterest
Email
Other Apps

Comments

Cheat Sheet on Classes, Traits and Interfaces on PHP/Hack

Haven't coded any PHP/Hack for a while, just recall my memory when I need to use Classes, Traits and Interfaces on PHP/Hack. :) Classes It is the blueprint of the Object. It has the states. Single inheritance Vs Multiple inheritance, the former one is easy to understand, but the latter one is more power but complex. Traits It is a group of similar functions It does not have a state. Multiple traits. Trait itself can have the implementation in the source code, it is just like copy & paste practice for code sharing. Of course, trait can have the abstract keyword in the methods to explicitly require the classes to implement those methods, like interface. Interfaces It is a group of functions that the Objects MUST implement. It does not have a state. Multiple interfaces. Interface itself cannot have the implementation in the source code as compared with the trait. Note: Abstract in PHP. Basically, it defines the signature of the classes / methods...

Matrix Operations in Machine Learning

One of the interesting things to play around machine learning is to optimize the computational cost in training the learning model through matrix operations. Taking a 3 layers neuron network as an example, it has one input layer, one hidden layer and one output layer, it is a typical model being used for recognizing handwritten numbers as shown below. Without using matrix operation, one can loop through K training samples to compute the output in layer 2 and 3, and finally get the summation of the total cost J. This code is more concise and easy to understand. J = 0; % J is the cost of the network for i = 1:K % K is the number of training samples % layer 1 a1 = X(i,:); % X is the input features % layer 2 z2 = Theta1 * a1'; % Theta1 a2 = sigmoid(z2); % sigmoid is logistic function % layer 3 z3 = Theta2 * [1; a2]; % Theta2 is a3 = sigmoid(z3); % yy is the output J...

Powered by Blogger

Safe JavaScript Templating Portal

Comments

Post a Comment