GAN (Generative Adversarial Networks)

One of the biggest and most fascinating examples of this and machine learning involves the word GAN and much of it revolves around the Nvidia company.

Now Nvidia was started in 1993 and has been at the forefront of creating advanced computer graphics cards since then.

Now they have these cards that are used for everything from advanced gaming to mining cyber currency to artificial intelligence applications.

Now the one project of interest that they have going is the Gau Gan project which is able to create original artwork that rivals the work of the masters through the use of mathematics and algorithms basically.

Basically its about inputting or training the software engine or the Nvidia model with thousands of images which is able to then process it all and create art. Yes we are missing a hundred details in this description 🙂

Lets describe what GAN is and then go back and talk about how it and machine learning is being used on websites now.

GAN stands for Generative Adversarial Networks and basically you would have two neural networks competing with each other to arrive at some synthetic data which could be a piece or art or music.

Now on the Internet there are many companies now where you can instantly create totally realistic human faces on the spot that you can buy to use in your company brochures or websites using this technique.

Some sights are and this persondoesnotexist where you land on the website and you get a different face every few moments created automatically.

On some of these sites you can create a dozen or so faces in a minute and then download and use them for free or pay perhaps $40 a year to use some of them or more.

The software called StyleGAN was develope by Nvidia and then released as part of the OpenSource movement so people can now incorporate it into their projects.

There is a neat graphics/video technology called Augmented learning which is starting to be used on Youtube.

It is not AI but actually a form of VR or virtual reality and it is amazing because a user can put an image of an elephant or a dinousaur beside them in the youtube video and it will look like that creature is actually in the room with that person.

This software called ARLOOPA allows you to create augmented learning examples in a matter of minutes.

So strap yourself in and get ready to learn about things like: machine learning, natural language processing, vision systems, facial recognition, AGM (artificial general intelligence), chatbots, tensorflow and other neural network libraries…. stereo pi, Google’s Coral Edge tpu accelerator, GPT on Pi4 using tensorflow, good ol’Dr Guero (Masahiko Yamaguchi),… tinman,