Application of the generative approach to creating artworks
by Bogdan Soban
The development of informational technologies gives a lot of new possibilities to use computer in other non standard fields of using. I have in mention a large researches on artificial intelligence where computer simulates human thinking, creating new solutions and resolving different problems in a way of intelligence data processing. One of this method is known as “generative art”, a wide area of research inside artificial creativity, an important part of artificial intelligence.
There are many definitions of creativity. Dictionaries give the following meanings: to cause to exist, bring to being, originate, to give rise to, produce, an original product of human invention or imagination, something characterized by originality and expressiveness, imaginative, compose, generative, innovative, originate etc. There are many aspects of creativity but one definition would include the ability to take existing objects and combine them in different ways for new purposes. A simple definition of the creativity is the action of combining previously uncombined elements. Creativity is the ability to generate novel and useful ideas and solution to everyday problems and challenges. Creativity is also the origination of new ideas that were previously non-existent. There is no one definition of creativity that everyone can agree with. Creativity researches, mostly from the field of psychology, usually claim that being creative means being novel and appropriate. Subsumed under the appropriateness criterion are qualities of fit, utility and value. New, unpredictable, unique and unrepeatable solutions are results of human mind, said also of brain processing in words o informational technology. Often great solutions are products of silly ideas. Silly ideas are those that do not seem to follow rules, and are consequently thrown out, but modification of these ideas can result in a different path to the desired objective through unusual connections.
Observing types of creativity there is a distinction between real-time creativity and multistage creativity. Real-time creativity is spur-of-the moment, improvisational, and demands output in a short interval of time and is property of very intelligent people. In multistage creativity, sufficient time is allowed for the generation and selection of ideas. Also creative thought can be divided in two types: divergent and convergent reasoning. Divergent thinking is the intellectual ability to think of many original, diverse, and elaborate ideas while convergent thinking is intellectual ability to logically evaluate, critique and choose the best ideas from a selection. Both abilities are required for successful creative output.
Artificial creativity arises, in part, out of the general field of artificial intelligence, but also out of the interest of a wide range of creative practitioners curious about the roles of emerging technology. It is clear that new technology can act as aid to creativity by providing new kinds of tools but there is a further question as to whether there is a prospect of machines competing with people in creative endeavor.
At the beginning some definition of artificial intelligence (AI) to understand better the meaning of artificial creativity. At first definition of an intellect: this is a thing that knows nothing but it can learn. The most widely spread definition of AI is the so called Turing's test. We place something behind a curtain and it speaks with us. If we can't make difference between it and human being then it will be AI. Speaking of AI and considering that every calculating device can be modelled by a program means that we are going to look for AI in the set of programs. Finally a formal definition: AI will be such a program which in an arbitrary world will cope not worse than a human.
And now a short analysis of what, in the most general terms, artificial creativity (AC) is. AC is a field that tries to determine in computational details how creativity works. But for AC to play this role we need to be able to understand creativity of different non-human kinds and believe that this is something that human-designed machines can achieve. If we can now construct successful models that accord with general account of creativity, then we can and should think of them as creative systems. Now is opening a very important question: how to make a computer creative if we know that computers can only blindly follow fixed rules defined in the program. We find the answer of this question in a special types of programmes called genetic algorithm.
The genetic algorithm is a model of machine learning, which derives its behavior from a metaphor of processes of evolution in nature. This is done by the creation within a machine of population of individuals (programs) represented by chromosomes, in essence a set of character strings that are analogous to the base-4 chromosomes that we see in our own DNA. The individuals in the population then go through processes of evolution using reproduction, selection, mutation and crossover operations.
A COMPUTABLE MODEL OF CREATIVITY
Human creativity, then, is a complex and poorly understood phenomenon and is quite difficult to define a general theoretical model to describe it. Even more remote it the possibility of a generalised computable model which can be used for writing programmes to simulate creativity. In a practice we use limited models described as a function of different elements. For application as architectural design, visual art, music, poetry etc we can take a model out of language-based view of art: artwork is a function of vocabulary, grammar, selection and interpretation. Depending on the sort of artwork to be produced, the vocabulary will comprise drawn, three-dimensional objects, letters, words, phrases, parts of buildings etc.
In order to determine the way in which the elements are to be interrelated to produce a new creation, a grammar is defined. The task of choosing which items of the vocabulary might potentially be used at any instant is the selection operation. And finally the interpretation takes into account the way in which the final result of the work are to be manifested.
Generative art (GA) is a new computer method for developing ideas and creating new solutions in al fields of creativity. It not refer only to plastic art but in the word take on experiments in architecture, town-planning, industrial design, music, poetry etc. this new approach copy evolution processes in nature including DNA code, the moment of birth, grow process until wanted grade of maturity. Actual there are genetic designed programmes which realize, as the nature does, an endless sequence of always different, unique and unpredictable solutions. The possibilities of generative art are still being explored by academic, creative and commercial sectors, but there seems to be a mutual understanding that this new form of creativity plays an important role in exploring new areas of art. An artwork produced with GA approach poses the “big question” of authorship. If, by automating one's creative process, one can produce an artwork automatically, where do we draw the boundary of creativity? Who is the artist, the coder or the executor? Does the artefact become irrelevant? Is the code (process) the artwork?
All these philosophical questions are examined and discussed at the International Conference of Generative Art, which take place on Milan Polytechnic, Italy, this year for the fourth time. A wide area of GA application is seen observing topics of the conference: artificial life, artificial intelligence, architecture projects, industrial design projects, visual art, generative music, poetry, visual grammar, design approach, virtual environment, entertainment, communication, generative robots etc. (www.generativeart.com)
MY EXPERIENCES ON GENERATIVE ART
My first artworks generated with generative approach were produced at the beginning of eightyths when C64 was available. Simulating creative process on the computer was a silly idea but it functioned. Simple programmes generated always new, unpredictable and never repeated graphic form on the screen. Operating with coloured points, lines and plains programmes created pure abstractions defined in its field as geometric abstractions. During the process of creation a computer was completely autonomous to the highest possible degree. By means of a coincidental function the computer itself was free to choose the largest possible number of elements determining a graphic image such as type, number, shape, size, colour, position on the screen etc.
The next step was programming stylised images of nature. Respecting the law of reality, computer was less free on choosing elements of the artwork. A typical example of stylised nature is Karst , a very colourful and poetic landscape especially in autumn. All these programmes use pragmatic instructions to define the rules by which such artworks are executed. It means that the final solution is predictable on the level of content but unpredictable on the level of appearance.
I made a quality step in my researches started developing generative programmes using mathematic formula on creating graphic images. Results were absolutely unpredictable and every run was a real surprise for me. Using two casually selected programmes I tested crossover operation typical for genetic programming.
Actually I am working on application of different variable elements produced by simulated kinematics systems as input in genetic algorithm. For example: a kinematics system composed of two circulating points in Cartesian coordinate system produces a lot of variables such a centres of rotation, radiuses, angle speeds, positions of the points, distances between different points, different angles, the distance between elaborated pixel and other points etc. Some variables are changing values continually depending of the time, some of them are assumed by current process. All these variables are input to genetic process, which creates something. And then evolving right mathematic formula on computing the colour of the pixel is possible to produce wonderful shapes. Developing these concept it could be a god idea to use as on-line input the noise of natural or artificial processes around us (waving of the sea, noise of the street etc). Some examples of my work is possible to see on my web site www.soban-art.com
All my programmes are written in GWBASIC programming language and they run in DOS environment without using any modern graphic tools or interfaces. Saving images needs to run programmes under DOS prompt inside WINDOWS using utilities for file transfer. It is a great satisfaction and challenge in the same time using an archaic programming language inside modern environment of WINDOWS98. I think to continue on this way because I know, there are no other influences to produced artworks as me in existing time and space, my program and my computer. There are no other programmes or tools, which might disturb the pure creative process of machine.