So far, we've looked at how artificial intelligence technology is being applied to creative activities in the literature and music sector. This time, we will look at artificial intelligence in the field of art, which is another creative field, and finally, we will look at the future of creative artificial intelligence.
See [AI by our side] Does artificial intelligence dream of artists (1)
See [AI by our side] Does artificial intelligence dream of artists (2)
In the film, <아이, 로봇> human detective Dell Spooner has the following conversation with the robot Sunny:
Spooner: Can robots write symphonies? Can a robot draw a great masterpiece on a canvas? (Can a robot write a symphony? Can a robot turn a canvas into a beautiful masterpiece?)
SUNNY: Can you do it? (Can you?)
It's been less than 20 years since this movie was released, but artificial intelligence for writing, composing, and drawing has already appeared. Nevertheless, it is a matter of humankind, and on one side of my mind, “Artificial intelligence, should I recognize you as an artist? I can't dismiss the question. It's an art I've always believed that only humans can do, but now technology has reached a level where it is difficult to clearly distinguish between works created by machines and humans.
So in the future, should humans take their hands off art and just appreciate machine-made art?
AI painter who paints abstractions
An image created by Deep Dream, Image source
A typical artificial intelligence artist is Deep Dream (Deep Dream), created by Google. It is said that Deep Dream recognizes and stores images through artificial neural networks, and reinterprets and visualizes the features of these images. And the result is dreamy and abstract, just like the name Deep Dream.
This principle of deep dreaming uses an algorithm to create abstract images by learning Van Gogh's style of painting. First, the algorithm splits the elements contained in the image, analyzes them, converts them into data, and finds patterns. The learned image pattern is then applied to create a new image. Thanks to this, even ordinary objects we see in our everyday lives can be turned into abstract art.
Google has a website so that new images can be created based on Deep Dream technology by directly uploading images directly (https://deepdreamgenerator.com) was left open. Now, anyone can use artificial intelligence to create their own pictures and enjoy them.
The artificial intelligence artist who revived Rembrandt
The Next Rembrandt, Image source
In the Netherlands, an artificial intelligence artist called The Next Rembrandt (The Next Rembrandt), who learned data from Rembrandt's works using deep learning, came out. It is the result of a project planned by advertising company Walter Thompson and collaborated with ING and Microsoft for 2 years.
What is particularly surprising is that it's not just about drawing Rembrandt-style paintings, but that it also perfectly reproduces the texture of the painting surface created by paint. Next Rembrandt analyzed the three-dimensional composition of the picture due to the thickness of the paint with data, built it with an algorithm, and succeeded in reproducing the unique texture of oil painting using a 3D printer based on this.
Artificial intelligence creates new styles
images created by CAN, Image source
Both Deep Dream and Next Rembrandt imitated traditional artists' styles of painting. Of course, even for humans, the beginning of creation begins by looking at and imitating numerous works of art.
So how do humans create new art? Colin Martindale (Colin Martindale), who has studied artistic creation, defines the following conditions for creating a new style: “First, it must not be classified as one of the existing styles (such as impressionism), and at the same time, it must have a form that can be called a work of art.” That's it.
Based on this theory, Facebook's AI team proposes CAN (Creative Adversarial Networks), which draws a new style of painting. Based on a generative adversarial network (Generative Adversarial Network), CAN used existing pictures and style classification information for each picture as learning data. In other words, after learning the works drawn by many existing artists, I succeeded in drawing a new style of painting.
It is said that CAN drew a new picture from an arbitrary vector image, but learned not to feel heterogeneity as a work of art without imitating an existing style. As a result, according to the experimental results published by the research team, most people have reached a level where they are unable to distinguish between traditional modern art and pictures drawn by artificial intelligence.
While finishing
Imitation is said to be the mother of creation. Since ancient times, humans have continued to expand the field of art by creating new forms of art and recreating them in various ways. Also, in this process, the creative techniques and tools used by artists have continued to evolve.
Until just a few decades ago, did humans imagine that computers would draw and make music? While looking at various examples, I began to think that the impact of artificial intelligence in the field of creation might also be in this context.
If so, wouldn't it be correct to think that artificial intelligence expands new possibilities rather than replacing existing artists or creative fields?
Just as tools such as brushes, paints, and musical instruments have appeared to help artists, artificial intelligence can also be another creative tool and even a partner for new collaborations.
References
[1] Deep Dream https://en.wikipedia.org/wiki/DeepDream
[2] Google's artificial intelligence 'Deep Dream' that also draws pictures https://www.news1.kr/articles/?2602024
[3] Rembrandt's new work drawn by artificial intelligence? http://www.techholic.co.kr/news/articleView.html?idxno=51918#rs
[4] The Next Rembrandt documentary https://youtu.be/IuygOYZ1Ngo
[5] Machine Creativity Beats Some Modern Art https://www.technologyreview.com/2017/06/30/150666/machine-creativity-beats-some-modern-art/
[6] [Asking the meaning of AI ART art] ⑧ Dr. Lee Kwang-hee “What is artistic creativity for AI painters” http://it.chosun.com/site/data/html_dir/2019/11/30/2019113000307.html
[7] [IT motto] Generative Adversarial Neural Network (GAN) http://www.bloter.net/archives/311614
[8] CAN: Creative Adversarial Networks Presents “Art” by Learning About Styles and Deviating from Style Norm https://arxiv.org/pdf/1706.07068.pdf
Good content to watch together
[AI by our side] Does artificial intelligence dream of artists (1)[AI by our side] Does artificial intelligence dream of artists (2)[AI by our side] Does artificial intelligence dream of artists (3)
