Setting up denoising autoencoders

Denoising autoencoders are a special kind of autoencoder with a focus on extracting robust features from the input dataset. Denoising autoencoders are similar to the previous model except with a major difference that the data is corrupted before training the network. Different approaches for corruption can be used such as masking, which induces random error into the data.

Getting ready

Let's use the CIFAR-10 image data to set up a denoising dataset:

• Download the CIFAR-10 dataset using the download_cifar_data function (covered in Chapter 3, Convolution Neural Network)
• TensorFlow installation in R and Python

How to do it...

We first need to read the dataset.

Reading the dataset

1. Load the CIFAR dataset using the steps explained in Chapter 3, Convolution Neural Network. The data files data_batch_1 and data_batch_2 are used to train. The data_batch_5 and test_batch files are used for validation and testing, respectively. The data can be flattened using the flat_data function...