Human-AI Generated Art Gallery
Designed by: Kexin LU, Hongxun JIN, Qijia LI
Over the past few years, generative machine learning and machine creativity have continued to grow and attract a broader audience to machine learning. New types of media creation are possible to be created, with AI to develop algorithms and techniques that endow computers with an understanding of our world.
Our Human-AI generated art gallery presenting a set of selected works to deliver the concept of a virtual memory space striking audiences’ emotional echo by filling in the gaps between fragments of common daily visions with AI-generated scenarios to build a sequence of coherent images where we implement the GAN latent space interpolation between random vectors from input sets of images as a creative way of generating non-existent memories, we can actually customize our own version of virtual memories about the evolvement of the world.
Background
Memory is a process of preserving and reconstructing our past experiences and knowledge, like neural connection, it bridges bits and pieces of our life together, making us who we are in the very present moment.
The traditional neural theory divides the memory process into three stages: encoding, storage, and retrieval; and our brain could make glitches to any of these three stages, so there is a chance that our memories are never 100% truth.
What if we take a second look at the familiar scenes around us, what memories would they recall? What if we try to remember with AI’s help?
From the classical Matrix trilogy to the latest popular series West World, the topic of artificial memory manipulation has been widely discussed and interpreted in various ways, what comes to us is that an artificial memory is often composed of some key visual elements while the manipulation changes whether the sequence of these fragments or creates fictional data in the blurry areas. And with the technique of StyleGAN where we implement the GAN latent space interpolation between random vectors from input sets of images as a creative way of generating non-existent memories, we can actually customize our own version of “memories” based on the input data sets and the controlling parameters.
Method
There are many technologies or phenomena in the field of AI that fascinate artists, from convolutional neural networks (CNN), or recurrent neural networks (RNN) for ordered data, to more complex and mixed models. Most of the AI artists who do graphics like GAN, and even ridiculous usages such as “GANism” appear. I personally feel that GAN’s sense of presence in the AI industry is not as strong, however, many artists have found something that resonates in GAN.
The media objects in the eyes of the computer are not the same as those in our human’s eyes. From a graphics point of view, a cup is composed of many pixels, each pixel contains four values, so the image of the cup can be converted into a string or a very long array. What the so-called hidden space / latent space is doing is to project a high-dimensional object like a cup into a relatively low-dimensional space in order to represent the change of real data in the low-dimensional space, or reaction training probability relationship of data. Although a lot of details will be discarded in this process, the computer has thus gained an understanding of the modalities and characteristics of the object.
Neural networks give artists a lot of inspiration only in concept, because it understands images and words in a very different way from humans. In the eyes of computers, images, text, and audio are different types of data. These data are all distributed in the space that can be operated on. Like the previous example, an artist may take the object as whether visual or emotional elements, but in the algorithm’s view, the image is a continuous sequence of values. The text is also similar, it can become a vector, and each feature of the vector is represented as a word appearing in the text, so when the word vector is unified into the same space, there is also the concept of distance.
Artificial intelligence (AI) is a unique and controversial form of technology because it represents another form of intelligence of our own. While human artists may take years training themselves from imitation to innovation, GAN could achieve the same point in hours. From there, artificial intelligence began to board the stage of the art-making process. What’s more, with the packed machine learning application, artists nowadays do not necessarily need to be able to understand the tedious complex math model and tuning parameter in code to generate a satisfying piece.
Even if the artist has an algorithm tool as an assistant, what data he uses, which model he uses, how to clean the data, how to structure, and how to select and display based on the results of the algorithm are still subjective. From this perspective, although the tool is effective, the effectiveness here is limited, after all, it cannot help the artist to do everything. Any creation will definitely encounter a lot of hardships or setbacks, and they will leave traces in the work. I believe that such traces will definitely be seen in the best works. In many cases, the process of feeding and feedback to the model is resuming. If the artist encounters into the situation where what he made does not meet the expectations, he then has to rework, select, and even overthrow it.
Precedents
DATA GATE _ World’s First AI Astronomical Research Data Sculpture Public Art by Ouchhh
The installation consists of 3 parts; Form, Light and Space. Light is world’s first artwork based upon the idea of utilization of Machine Learning in the context of space discovery and astronomical research through NASA’s Kepler Data Sets.
By using the Kepler data from NASA, the public will be able to observe the exoplanets [planets that orbit around other stars] which human life can exist in. By taking this concept one step further, Ouchhh aims for this artwork to be considered as a gate between our planet and other habitable planets around the universe.
Ouchhh visualized and stylized the findings of these Neural Networks for identifying exoplanets using the dimming of the flux. The resulting work will invite visitors to plunge into the fascinating world of space discovery through immersive data sculpture. The installation will offer a poetic sensory experience and is meant to become a monument of mankind’s contemplative curiosity and profound need for exploration.
resource: http://ouchhh.tv/
Universe of Water Particles on a Rock where People Gather
Crossing the boundaries of other works, sometimes influencing other works, water flows from the Athletics Forest, Graffiti Nature — High Mountains and Deep Valleys, and accumulates into the waterfall of the Universe of Water Particles on a Rock where People Gather.
The Rock where People Gather is reproduced in a virtual three-dimensional space. Water is simulated to fall onto the rock, and the flow of the water draws the shape of the waterfall. The water is represented by a continuum of numerous water particles and the interaction between the particles is then calculated. Lines are drawn in relation to the behavior of the water particles. The lines are then “flattened” using what teamLab considers to be ultra subjective space.
When a person stands on the rock or touches the waterfall, they too become like a rock that changes the flow of water. The flow of water continues to transform in real-time due to the interaction of people. Previous visual states can never be replicated, and will never reoccur.
resource: https://borderless.teamlab.art/ew/iwa-waterparticles/
Material
Our gallery consists of 6 pieces of artwork. We trying to use AI to dig the unseen aspects of things from large scale as nebulas to small scale as species.
Nebula
The universe is a living body and Space has its own pulse and rhythm. Thanks to the new technology, we can ‘see’ the sound of the universe, the pulse and breath of the night, and the hurried trail of countless dreams in this silence following the rhythm in the video.
Resource: RunWay ML
Landscape
Change, the only unchangeable truth to the world, applies to plate tectonics as well. The land on earth is divided into seven continents by vast waters, and everything becomes more vivid in the depths of the ocean that we cannot see. Material on the earth is destroying itself, melting itself, and then reshaping. A dozen cold, hard plates slid slowly over the hot mantle inside the earth, subducting each other, and occasionally colliding, in an invisible way.
Ocean
Since ancient times, the ocean has been closely related to human production and life. Ocean, those large beautiful bodies of water are home to an infinite number of strange creatures and bewildering phenomena. However, 95% of the ocean is unexplored, unknown, unseen by human eyes. Following the video, let’s quietly unlock the corner of this mysterious veil.
Nature
Mountain, land, river, habitat, all the things on the planet, over geological timescales, will change in volume, shape, temperature, and chemistry, some waxing larger, others drying out or freezing as water moves between planetary interior, crustal surface, atmosphere, and outer space. The planet is telling a story itself in its own way.
Architecture
Architecture, an ancient art form that mostly related to human being’s development, has undergone diverse stages in different societal contexts and historical backgrounds: Baroque, Roman, Gothic, Modern, Contemporary. We consider buildings as vectors of a cultural significance, that express through their geometry, taxonomy, typology, and decoration a certain style. Past memory emerges with the flashback of the old buildings, while new inspiration is approaching towards tomorrow with new AI solutions that will serve as an intersection between context and new tech.
Species
Butterfly evolution creates the diversification of butterflies through geologic time and over a large portion of the Earth’s surface. Those universally loved flying works of art offer many reasons to celebrate design in nature in an unseen way for human beings. Tracing back to ancient times, we utilize the AI tool to experience the special aesthetic beauty beyond the requirements of survival of this species.
Overview
All of the scenes in this art gallery are set up to showcase the familiar entities in an unexplored status by representing AI-generated scenarios. The exhibition is designed for stimulating the intrinsic imagination for the entities that the visitors take for granted for years to customize their own versions of virtual memories about the evolvement of the world. It represents the nature of entities in deconstructive ways to unveil the unseen forms of existence of those memories and reconstruct the past experiences and knowledge then provide a chance to for us to reflect and take a second look at the familiar scenes around us.
