Deep Learning with Tensorflow - Autoencoders with TensorFlow Video Lecture | Tensorflow: Learning made Easy - AI & ML

Hello, and welcome!
In this video, we're going to show you how
to implement an autoencoder using the TensorFlow
library.
We will be using the MNIST dataset of handwritten
digits.
Let's run the imports so that the MNIST dataset
is available to use.
Now, let's specify the parameters that are
going to be used by our Neural Network.
Next we need to create our encoder.
For this, we are going to use sigmoidal functions.
This type of function has a derivative that
is well-suited for backpropagation.
Afterwards we specify the decoder.
You can see that 'layer_1' in the encoder
is 'layer_2' in the decoder, and vice-versa.
Let's construct our model.
In the 'cost' variable, we have the 'loss'
function.
In the 'optimizer' variable, we have our gradient
used for backpropagation.
Now let's execute the training process.
Note that running for 20 epochs will take
some time.
Now, let's apply encode and decode for our
tests.
And let's visualize our graphs!
As you can see, the reconstructions were successful,
though there was some noise added to the image.
By now, you should understand how to implement
and train an autoencoder using the TensorFlow
library.
Thank you for watching this video.