吴恩达老师机器学习ex4.Neural Network Learning

吴恩达 机器学习 第五周作业 Neural Network Learning

Octave代码

nnCostFunction.m

function [J grad] = nnCostFunction(nn_params, ...
                                   input_layer_size, ...
                                   hidden_layer_size, ...
                                   num_labels, ...
                                   X, y, lambda)
%NNCOSTFUNCTION Implements the neural network cost function for a two layer
%neural network which performs classification
%   [J grad] = NNCOSTFUNCTON(nn_params, hidden_layer_size, num_labels, ...
%   X, y, lambda) computes the cost and gradient of the neural network. The
%   parameters for the neural network are "unrolled" into the vector
%   nn_params and need to be converted back into the weight matrices. 
% 
%   The returned parameter grad should be a "unrolled" vector of the
%   partial derivatives of the neural network.
%

% Reshape nn_params back into the parameters Theta1 and Theta2, the weight matrices
% for our 2 layer neural network
Theta1 = reshape(nn_params(1:hidden_layer_size * (input_layer_size + 1)), ...
                 hidden_layer_size, (input_layer_size + 1));

Theta2 = reshape(nn_params((1 + (hidden_layer_size * (input_layer_size + 1))):end), ...
                 num_labels, (hidden_layer_size + 1));

% Setup some useful variables
m = size(X, 1);
         
% You need to return the following variables correctly 
J = 0;
Theta1_grad = zeros(size(Theta1));
Theta2_grad = zeros(size(Theta2));

% ====================== YOUR CODE HERE ======================
% Instructions: You should complete the code by working through the
%               following parts.
%
% Part 1: Feedforward the neural network and return the cost in the
%         variable J. After implementing Part 1, you can verify that your
%         cost function computation is correct by verifying the cost
%         computed in ex4.m
%
% Part 2: Implement the backpropagation algorithm to compute the gradients
%         Theta1_grad and Theta2_grad. You should return the partial derivatives of
%         the cost function with respect to Theta1 and Theta2 in Theta1_grad and
%         Theta2_grad, respectively. After implementing Part 2, you can check
%         that your implementation is correct by running checkNNGradients
%
%         Note: The vector y passed into the function is a vector of labels
%               containing values from 1..K. You need to map this vector into a 
%               binary vector of 1's and 0's to be used with the neural network
%               cost function.
%
%         Hint: We recommend implementing backpropagation using a for-loop
%               over the training examples if you are implementing it for the 
%               first time.
%
% Part 3: Implement regularization with the cost function and gradients.
%
%         Hint: You can implement this around the code for
%               backpropagation. That is, you can compute the gradients for
%               the regularization separately and then add them to Theta1_grad
%               and Theta2_grad from Part 2.
%

% Part1
X = [ones(m, 1) X];

for i = 1 : m;

z2 = Theta1 * X(i, :)';
a2 = sigmoid(z2);
a2 = [ones(1, 1); a2];
z3 = Theta2 * a2;
a3 = sigmoid(z3);

temp = y(i);
y_cls = zeros(num_labels, 1);
y_cls(y(i), 1) = 1;

J = J + 1 / m * sum(-y_cls .* log(a3) - (1 - y_cls) .* log(1 - a3));

end;

J = J + 0.5 * lambda / m * (sum(sum(Theta1(:, 2:end) .^ 2)) + sum(sum(Theta2(:, 2:end) .^ 2)));

% Part2
for t = 1 : m;
a1 = X(t, :)';
z2 = Theta1 * a1;
a2 = sigmoid(z2);
a2 = [ones(1, 1); a2];
z3 = Theta2 * a2;
a3 = sigmoid(z3);

temp = y(t);
y_cls = zeros(num_labels, 1);
y_cls(y(t), 1) = 1;
d3 = a3 - y_cls;

d2 = Theta2'(2:end, :) * d3 .* sigmoidGradient(z2);

Theta1_grad = Theta1_grad + d2 * a1';
Theta2_grad = Theta2_grad + d3 * a2';

end;

% Part 3
Theta1(:, 1) = 0;
Theta2(:, 1) = 0;
Theta1_grad = 1 / m * Theta1_grad + lambda / m * Theta1;
Theta2_grad = 1 / m * Theta2_grad + lambda / m * Theta2;

% -------------------------------------------------------------

% =========================================================================

% Unroll gradients
grad = [Theta1_grad(:) ; Theta2_grad(:)];


end

sigmoidGradient.m

function g = sigmoidGradient(z)
%SIGMOIDGRADIENT returns the gradient of the sigmoid function
%evaluated at z
%   g = SIGMOIDGRADIENT(z) computes the gradient of the sigmoid function
%   evaluated at z. This should work regardless if z is a matrix or a
%   vector. In particular, if z is a vector or matrix, you should return
%   the gradient for each element.

g = zeros(size(z));

% ====================== YOUR CODE HERE ======================
% Instructions: Compute the gradient of the sigmoid function evaluated at
%               each value of z (z can be a matrix, vector or scalar).


g_temp = 1.0 ./ (1.0 + exp(-z));
g = g_temp .* (1 - g_temp);



% =============================================================




end

运行结果

Octave command prompt

octave:20> ex4

Loading and Visualizing Data ...
Program paused. Press enter to continue.


Loading Saved Neural Network Parameters ...

Feedforward Using Neural Network ...
Cost at parameters (loaded from ex4weights): 0.287629 
(this value should be about 0.287629)

Program paused. Press enter to continue.


Checking Cost Function (w/ Regularization) ... 
Cost at parameters (loaded from ex4weights): 0.383770 
(this value should be about 0.383770)
Program paused. Press enter to continue.


Evaluating sigmoid gradient...
Sigmoid gradient evaluated at [-1 -0.5 0 0.5 1]:
  0.196612 0.235004 0.250000 0.235004 0.196612 

Program paused. Press enter to continue.


Initializing Neural Network Parameters ...

Checking Backpropagation... 
  -9.2783e-03  -9.2783e-03
   8.8991e-03   8.8991e-03
  -8.3601e-03  -8.3601e-03
   7.6281e-03   7.6281e-03
  -6.7480e-03  -6.7480e-03
  -3.0498e-06  -3.0498e-06
   1.4287e-05   1.4287e-05
  -2.5938e-05  -2.5938e-05
   3.6988e-05   3.6988e-05
  -4.6876e-05  -4.6876e-05
  -1.7506e-04  -1.7506e-04
   2.3315e-04   2.3315e-04
  -2.8747e-04  -2.8747e-04
   3.3532e-04   3.3532e-04
  -3.7622e-04  -3.7622e-04
  -9.6266e-05  -9.6266e-05
   1.1798e-04   1.1798e-04
  -1.3715e-04  -1.3715e-04
   1.5325e-04   1.5325e-04
  -1.6656e-04  -1.6656e-04
   3.1454e-01   3.1454e-01
   1.1106e-01   1.1106e-01
   9.7401e-02   9.7401e-02
   1.6409e-01   1.6409e-01
   5.7574e-02   5.7574e-02
   5.0458e-02   5.0458e-02
   1.6457e-01   1.6457e-01
   5.7787e-02   5.7787e-02
   5.0753e-02   5.0753e-02
   1.5834e-01   1.5834e-01
   5.5924e-02   5.5924e-02
   4.9162e-02   4.9162e-02
   1.5113e-01   1.5113e-01
   5.3697e-02   5.3697e-02
   4.7146e-02   4.7146e-02
   1.4957e-01   1.4957e-01
   5.3154e-02   5.3154e-02
   4.6560e-02   4.6560e-02
The above two columns you get should be very similar.
(Left-Your Numerical Gradient, Right-Analytical Gradient)

If your backpropagation implementation is correct, then 
the relative difference will be small (less than 1e-9). 

Relative Difference: 2.08374e-11

Program paused. Press enter to continue.


Checking Backpropagation (w/ Regularization) ... 
  -9.2783e-03  -9.2783e-03
   8.8991e-03   8.8991e-03
  -8.3601e-03  -8.3601e-03
   7.6281e-03   7.6281e-03
  -6.7480e-03  -6.7480e-03
  -1.6768e-02  -1.6768e-02
   3.9433e-02   3.9433e-02
   5.9336e-02   5.9336e-02
   2.4764e-02   2.4764e-02
  -3.2688e-02  -3.2688e-02
  -6.0174e-02  -6.0174e-02
  -3.1961e-02  -3.1961e-02
   2.4923e-02   2.4923e-02
   5.9772e-02   5.9772e-02
   3.8641e-02   3.8641e-02
  -1.7370e-02  -1.7370e-02
  -5.7566e-02  -5.7566e-02
  -4.5196e-02  -4.5196e-02
   9.1459e-03   9.1459e-03
   5.4610e-02   5.4610e-02
   3.1454e-01   3.1454e-01
   1.1106e-01   1.1106e-01
   9.7401e-02   9.7401e-02
   1.1868e-01   1.1868e-01
   3.8193e-05   3.8193e-05
   3.3693e-02   3.3693e-02
   2.0399e-01   2.0399e-01
   1.1715e-01   1.1715e-01
   7.5480e-02   7.5480e-02
   1.2570e-01   1.2570e-01
  -4.0759e-03  -4.0759e-03
   1.6968e-02   1.6968e-02
   1.7634e-01   1.7634e-01
   1.1313e-01   1.1313e-01
   8.6163e-02   8.6163e-02
   1.3229e-01   1.3229e-01
  -4.5296e-03  -4.5296e-03
   1.5005e-03   1.5005e-03
The above two columns you get should be very similar.
(Left-Your Numerical Gradient, Right-Analytical Gradient)

If your backpropagation implementation is correct, then 
the relative difference will be small (less than 1e-9). 

Relative Difference: 2.01903e-11


Cost at (fixed) debugging parameters (w/ lambda = 3.000000): 0.576051 
(for lambda = 3, this value should be about 0.576051)

Program paused. Press enter to continue.


Training Neural Network... 
Iteration    50 | Cost: 4.366683e-01
Program paused. Press enter to continue.


Visualizing Neural Network... 

Program paused. Press enter to continue.


Training Set Accuracy: 96.060000