Wanna know more about data science? Make sure to check out my events and my webinar What it's like to be a data scientist and What’s the best way to become a data scientist !
AI for creative purposes
Artificial Intelligence has to be the hottest buzzword right now. A Google search at the time of writing returns more than 600 million results. A huge number of companies are working on AI products, or are using AI products. While in most cases what is termed artificial intelligence, is nothing else than good old-fashioned machine learning, it is clear that some of the things we are witnessing start reminiscing of real intelligence.
Out of all this, something that steals the spotlight is the use of AI for creative purposes. This has been a very recent trend which came with the advent of generative adversarial networks.
What are generative adversarial networks?
Generative adversarial networks are a type of deep neural network where two networks are playing a game against each other. One network, called the generator, generates artificial objects (e.g. images). The other network (called the discriminator) tries to figure out whether the image is a fake or not. This constant game ends up training a generator that can generate realistic images.
How realistic? Well the results speak for themselves. Check out the latest state of the art created by Nvidia. What if I told you all the images are of fake people?
So, the question as to what are the limits of AI’s creativity might have sounded a bit like an oxymoron a few years ago. Creativity is a unique human property. Art is a true form of expression. How can a machine be creative? However, GANs and other developments in AI make it hard to deny that AI can be creative.
OpenAI created a language generator that is so realistic that they decided to not release to the public in fear of it being used to generate fake news. The pop artist Taryn Southern used AI in order to generate music for her new album. Services like aiva and amper music can be used to generate music clips. Some have taken it a step further. The painting below was sold for over $400k by Christie’s in 2018.
Or check out this video of an algorithm that produces ever changing portraits.
I believe that AI has made us question again what is art and what is intelligence. And as a data scientist I decided I don’t want to just stand by and watch. I have decided to get personally involved in this exciting space. I envision a world where machines and humans can work in tandem, and I believe this time is now.
The deep neural network revolution in art
I am researching the application use of deep neural networks (generative or otherwise) for creative purposes. I have decided to partner up with Jack Oat in order to do that. For those of you who don’t know him, Jack Oat, usually goes by the name Mr London Boudoir, and is one most famous underground photographers and artists in London. Our goal is to raise money in order to take his portraits and use them as an input for work that will explore the horizon of AI and creativity. The project is called Face Off: An artificially intelligent introspective. We have used the GAN with close to 2000 photos in order to generate
You can read more about our plan and help our project on Patreon. We would be delighted if you support our independent work. Also, in order to support all new technologies, we decided to secure ownership of the digital assets on the Ethereum blockchain. By donating you can not only secure ownership of the work of art, but you can also add an encrypted message (e.g. your signature or a phrase) on the art, therefore making yours forever.
There will be a free exhibition taking place 11th of July until the end of the month, so feel free to visit us!