MAterail Ref. added

MelJan · Apr 22, 2017 · 6297f44 · 6297f44
1 parent 568499d
commit 6297f44
Show file tree

Hide file tree

Showing 6 changed files with 38 additions and 4 deletions.
diff --git a/docs/tutorials/Binary_RBM_MNIST_small.rst b/docs/tutorials/Binary_RBM_MNIST_small.rst
@@ -6,6 +6,25 @@ The model is small enough to calculate the exact Log-Likelihood and annealed imp
 
 For an analysis of advantage of centering in RBMs see `How to Center Deep Boltzmann Machines. Melchior, J., Fischer, A., & Wiskott, L.. (2016). Journal of Machine Learning Research, 17(99), 1–61. <http://jmlr.org/papers/v17/14-237.html>`_
 
+If you are new on RBMs, a good theoretical introduction is given by `Course Material ICA <https://www.ini.rub.de/PEOPLE/wiskott/Teaching/Material/index.html>`_ and in the following video.
+
+.. raw:: html
+
+    <div style="margin-top:10px;">
+      <iframe width="560" height="315" src="http://www.youtube.com/embed/bMaITeXhOaE" frameborder="0" allowfullscreen></iframe>
+    </div>
+
+and
+
+.. raw:: html
+
+    <div style="margin-top:10px;">
+      <iframe width="560" height="315" src="http://www.youtube.com/embed/nyk5XUklb5M" frameborder="0" allowfullscreen></iframe>
+    </div>
+
+Results
+***********
+
 The code_ given below produces the following output.
 
 Learned filters of a centered binary RBM on the MNIST dataset.

diff --git a/docs/tutorials/ICA_2D_example.rst b/docs/tutorials/ICA_2D_example.rst
@@ -6,7 +6,7 @@ Independent Component Analysis on a 2D example.
 Example for Independent Component Analysis (`ICA <https://en.wikipedia.org/wiki/Principal_component_analysis>`_) used
 for a blind source separation on a linear 2D mixture.
 
-If you are new on blind source separation, a good theoretical introduction is given in the following video.
+If you are new on ICA and blind source separation, a good theoretical introduction is given by `Course Material ICA <https://www.ini.rub.de/PEOPLE/wiskott/Teaching/Material/index.html>`_ and in the following videos.
 
 .. raw:: html
 
@@ -22,6 +22,9 @@ and the follow up video introduces to ICA.
       <iframe width="560" height="315" src="http://www.youtube.com/embed/ugiMhRbFnTo" frameborder="0" allowfullscreen></iframe>
     </div>
 
+Results
+***********
+
 The code_ given below produces the following output.
 
 Visualization of the data and true mixing matrix projected to the whitened space.

diff --git a/docs/tutorials/ICA_natural_images.rst b/docs/tutorials/ICA_natural_images.rst
@@ -8,6 +8,9 @@ on natural image patches. The independent components (columns of the ICA project
 
 See  `ICA_2D_example <ICA_2D_example.html#ICA_2D_example>`__ for a theoretical introduction.
 
+Results
+***********
+
 The code_ given below produces the following output.
 
 Visualization of 100 examples of the gray scale natural image dataset.

diff --git a/docs/tutorials/PCA_2D_example.rst b/docs/tutorials/PCA_2D_example.rst
@@ -5,14 +5,18 @@ Principal Component Analysis 2D example.
 
 Example for Principal Component Analysis (`PCA <https://en.wikipedia.org/wiki/Principal_component_analysis>`_) on a linear 2D mixture.
 
-If you are new on PCA, a good theoretical introduction is given in the following video.
+
+If you are new on PCA, a good theoretical introduction is given by `Course Material ICA <https://www.ini.rub.de/PEOPLE/wiskott/Teaching/Material/index.html>`_ and in the following videos.
 
 .. raw:: html
 
     <div style="margin-top:10px;">
       <iframe width="560" height="315" src="http://www.youtube.com/embed/9H-1FH1gn6w" frameborder="0" allowfullscreen></iframe>
     </div>
 
+Results
+***********
+
 The code_ given below produces the following output.
 
 The data is plotted with the extracted principal components.

diff --git a/docs/tutorials/PCA_eigenfaces.rst b/docs/tutorials/PCA_eigenfaces.rst
@@ -7,6 +7,9 @@ Example for Principal Component Analysis (PCA) on face images also known as `Eig
 
 See  `PCA_2D_example <PCA_2D_example.html#PCA_2D_example>`__ for a theoretical introduction.
 
+Results
+***********
+
 The code_ given below produces the following output.
 
 Some examples of the face images of the olivetti face dataset.

diff --git a/examples/RBM_binary_MNIST_small.py b/examples/RBM_binary_MNIST_small.py
@@ -122,12 +122,14 @@
 
 # Calculate true partition function
 logZ = estimator.partition_function_factorize_h(rbm, batchsize_exponent=h1, status=False)
-print("True Partition: {} (LL: {})".format(logZ, numx.mean(estimator.log_likelihood_v(rbm, logZ, train_data))))
+print("True Partition: {} (LL: {})".format(logZ, numx.mean(
+    estimator.log_likelihood_v(rbm, logZ, train_data))))
 
 # Approximate partition function by AIS (tends to overestimate)
 logZ_approx_ = estimator.annealed_importance_sampling(rbm)[0]
 print(
-"AIS Partition: {} (LL: {})".format(logZ_approx_, numx.mean(estimator.log_likelihood_v(rbm, logZ_approx_, train_data))))
+"AIS Partition: {} (LL: {})".format(logZ_approx_, numx.mean(
+    estimator.log_likelihood_v(rbm, logZ_approx_, train_data))))
 
 # Approximate partition function by reverse AIS (tends to underestimate)
 logZ_approx_up = estimator.reverse_annealed_importance_sampling(rbm, data=train_data)[0]