AutoEncoders Clearly Explained | With examples and theory

This video explains how AutoEncoders works in Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL). Firstly, AutoEncoders take a significant role in efficient data representation and come-up with a series of benefits in Data Science, including PCA (dimensionality reduction), Generative Adversarial Networks (GANs), data reconstruction tasks, data compression, and much more. It depend on you project and tasks in Data Science. With this video I explain the main idea, how AutoEncoders works, what is the intuition behind with some coding with Keras and Tensorflow. The Python code itself is similar like Multi-Layer Perception - MLP). In the coding part, there is a quick example how AutoEncoders can be applied for implementing PCA algorithm. And also, to other data science problems. In order to better illustrate the benefits I took case of chess game. Did you know that expert chess players were able to memorize the positions of all the pieces in a game by looking at the board for just five seconds, that is for most people is impossible. There AutoEncoders take places a role also, because it works similar like real human brains. An autoencoders looks at the inputs, convert them to an efficiency latent representation, and then spits out something that looks very close to the inputs. An autoencoder is always composed of two parts: an encoder (or recognition network) that converts the inputs to a latent representation, followed by a decoder (or generative network) that converts the internal data representation to the outputs. The outputs are often called the reconstructions because the autoencoders tries to reconstruct the inputs, and the cost function contains a reconstruction loss that penalizes the model when the reconstructions are different from the inputs. The most important keywords are highlighted and explained in the video with screenshots of outside references. The video explains a basic type of autoencoders in details only (example of undercomplete autoencoder). Other types of autoencoders such as stacked autoencoders, convolutional autoencoders, recurrent autoencoders, denoising autoencoders, sparse autoencoders, and variational autoencoders - are not explained. This could be the inspirations for the next videos. If you wish to get some materials on that, let me know in the comments below. I found useful with preparation of this video from the ML book written by Aurelien Geron (Hands-On Machine Learning with Scikit-Learn, Keras and Tensorflow, from O’Reilly). I tried to create the simplest explanation of autoencoders as possible. Like it if you liked it! Subscribe this channel to get more useful material in future. #autoencoders #gans #tensorflow