Preface
Art is the Queen of all sciences
communicating knowledge to all the generations of the world.
Leonardo da Vinci
Artistic behavior is one of the most valued
qualities of the human mind. Although artistic manifestations
vary from culture to culture, dedication to artistic tasks is
common to all. In other words, artistic behavior is a universal
trait of the human species.
The current, Western definition of art is
relatively new. However, a dedication to artistic endeavors
--- such as the embellishment of tools, body ornamentation,
or gathering of unusual, arguably aesthetic, objects --- can
be traced back to the origins of humanity. That is, art is ever-present
in human history and prehistory.
Art and science share a long and enduring
relationship. The best-known example of the exploration of this
relationship is probably the work of Leonardo da Vinci. Somewhere
in the 19th century art and science grew apart, but the cross-transfer
of concepts between the two domains continued to exist. Currently,
albeit the need for specialization, there is a growing interest
in the exploration of the connections between art and science.
Focusing on computer science, it is interesting
to notice that early pioneers of this discipline such as Ada
Byron and Alan Turing showed an interest in using computational
devices for art-making purposes. Oddly, in spite of this early
interest and the ubiquity of art, it has received relatively
little attention from the computer science community in general,
and, more surprisingly, from the artificial intelligence community.
In the initial years of artificial intelligence
research the main source of inspiration was human intelligence.
Recently, this traditional, somewhat anthropocentric, view of
intelligence has given rise to the search for other potential
sources of inspiration. There is a growing interest in biology-inspired
computing techniques, a broad area of research that incorporates
techniques such as evolutionary computation, swarm intelligence,
ant colony optimization, and artificial life. These techniques
offer a wide range of solutions and opportunities, for scientists,
who have always made an effort to understand and model nature,
and for artists, who have always used nature as a source of
inspiration. The use of a metaphor that is relevant for scientists
and artists helps to bridge the gap between the scientific and
artistic communities, and fosters the collaboration and transfer
of knowledge between the two domains.
In this line of thought, the seminal works
of Richard Dawkins, Karl Sims and William Latham led to the
emergence of a new research area, usually called Evolutionary
Art and Music, which is characterized by the use of nature-inspired
computation in artistic domains.
The early books edited by Peter Bentley and
David Corne gave evolutionary art some important exposure. Over
time, the growing interest in the area led to the appearance
of dedicated scientific events and special issues, fostering
the development of a strong research community and playing an
important role in the establishment of evolutionary art and
music as a meaningful research field. The current vitality of
the area is reflected in the existence of dedicated annual workshops
and special tracks at some of the main evolutionary computation
conferences (e.g., Evolved Art and Music and Evolutionary
Design at the IEEE Congress on Evolutionary Computation,
and EvoMUSART, the Evo* Workshop on Evolutionary Music and Art).
This thriving area of research is arguably at the verge of adulthood.
Its current stage of development calls for a book that (1) provides
a broad and coherent coverage of the field, (2) provides the
necessary background information for newcomers, and (3) establishes
directions for future research, thus providing a solid basis
for its further development. These are the main objectives of
the present book.
The book is aimed at a wide audience, including
researchers and artists, beginners and experts in the field,
and especially those who wish to explore the relationships between
nature, science and art. We consider that it is important to
shorten the gap between the scientific and artistic communities.
Hopefully this book is a step in that direction, and this concern
is reflected in the contents and structure of the book.
The book is divided into five parts: Evolutionary
Art, Evolutionary Music, Real-World Applications,
Artistic Perspectives, and Future Perspectives.
The first two parts of this book include some
of the most interesting works on the application of evolutionary
computation techniques in the fields of visual art, video, design
(Part 1), sound, music and performance (Part 2). Although these
chapters are mainly scientifically oriented, they all make relevant
artistic contributions.
The first chapter, by Matthew Lewis, provides
a thorough, and much needed, analysis of the state of the art
in the fields of evolutionary art and design, introducing key
concepts and terminology, reviewing nearly 200 publications,
describing the most prominent approaches, and identifying some
of the most relevant research topics in the area.
In the second chapter John Collomosse describes the use of evolutionary
computation techniques in the context of the non-photorealistic,
painterly, rendering of images. Starting with an overview of
artistic stylization algorithms, he then discusses the use of
genetic algorithms to increase control over the level of detail
in painting, and to enhance the usability of painterly rendering
algorithms. The closing chapter of the first part of the book
presents the ``Electric Sheep'' project, one of the largest
and longest ongoing evolutionary art experiments, involving
over 40,000 computers and people mediated using a genetic algorithm.
Scott Draves offers a description of the representation, genotype--phenotype
mapping and genetic operators that allow the evolution of fractal
flames movies and still images, and he then focuses on the long-term
behavior of the distributed system.
The fourth chapter takes us to the area of
evolutionary sound synthesis. James McDermott, Niall J.L. Griffith
and Michael O'Neill survey previous work in the area, and then
focus on the problem of automatically matching a target sound
using a given synthesizer, which involves building fitness functions
that take into account timbral, perceptual, and statistical
sound attributes. They report and thoroughly analyze the results
attained in a comprehensive set of experiments aimed to determine
the best combination of algorithm, parameters and fitness functions
for this problem, drawing conclusions and indicating future
work. Tim Blackwell describes the use of swarm intelligence
and granular synthesis techniques for the generation of novel
sounds, outlining the theoretical foundations of these techniques
and the practical aspects involved in their usage. The explanation
is illustrated by the detailed description of two swarm granulation
systems, Swarm Granulator and Swarm Techtiles,
and by the analysis of their behavior. In the sixth chapter,
which concludes the Evolutionary Music part of the book, Rafael
Ramirez, Amaury Hazan, Jordi Mariné; and Xavier Serra tackle
a challenging problem in computer music, producing an expressive
performance of a musical piece. They use a genetic algorithm
to build a computational model of expressive performance from
a set of examples of jazz saxophone performances. Later, they
use this model to automatically create performances of musical
pieces.
The third part of the book comprises chapters
that are characterized by the use of evolutionary art approaches
for real-world applications, providing valuable case studies.
Christian Jacob and Gerald Hushlak describe the use of evolutionary
and swarm design techniques in art, music and design, showing
how interactive breeding techniques can facilitate the creative
processes, and presenting a wide variety of examples in areas
that range from furniture design to swarm choreographies. In
the eighth chapter, Martin Hemberg, Una-May O'Reilly, Achim
Menges, Katrin Jonas, Michel da Costa Gonçalves and Steven
R. Fuchs take us to the domain of architecture, describing Genr8
--- an evolutionary system that allows the evolution of surfaces
generated through an organic growth algorithm --- and reporting
its use on six different architectural projects. Charlie D.
Frowd and Peter J.B. Hancock describe EvoFIT, a system
that allows the evolution of photorealistic human faces, and
explore its use for the production of facial composites of criminals.
Later, the artistic potential of EvoFIT is also analyzed,
and other potential application areas discussed. In the tenth
chapter, A.E. Eiben describes the modeling of the artistic styles
of the famous Dutch painters Piet Mondriaan and M.C. Escher,
giving particular emphasis to the mathematical modeling of Escher's
tilings and to the construction of an evolutionary system that
allows their generation.
One of the difficulties inherent to evolutionary
art and music is the difference between the scientific and artistic
perspectives. To lessen this problem the fourth part of this
volume gives voice to artists who employ or analyze biology-inspired
mechanisms in an artistic context. As such, it consists of chapters
where the artistic perspective is the most fundamental. The
interest of Nicolas Monmarché, Isabelle Mahnich and Mohamed
Slimane in swarm intelligence, ant colony algorithms and self-organization
leads them to an exploration of the artistic potential of these
concepts for the creation of spatio-temporal structures, which
is illustrated by the evolution of musical pieces and paintings.
In the twelfth chapter, Gunter Bachelier describes the three
levels --- basic, methodical and superordinate --- of his art
practice. His unique evolutionary art approach --- which relies
on a pixel-based representation, on the exchange of regions
of interest, and on the application of transformations to these
regions --- is thoroughly described. Later he presents his novel
evolutionary art approach, which also integrates aspects such
as multi-sexual reproduction and image templates, and ontogenetic
concepts such as spores or fruits. In the thirteenth chapter,
Jeffrey J. Ventrella presents Musical Gene Pool, an
application that allows the evolution of liquid music,
i.e., nonlinear music whose structure is continually able to
flow and rearrange, allowing serendipity. Alan Dorin presents
a survey of the use of virtual ecosystem simulation in the context
of generative electronic art. Based on a thorough analysis of
these systems, he concludes that their major strengths lie in
the ability to display multi-scaled complexity and to produce
novelty, and that their major weakness lies in their unpredictable
response to perturbation; he later describes methods to overcome
this weakness. In the concluding chapter of this part, Philip
Galanter, following the modernist tradition of the art manifesto,
proposes a new art approach, entitled Complexism, which
relies on the ``application of a scientific understanding of
complex systems to the subject matter of the arts and humanities''.
He compares it with modernist and postmodernist movements, arguing
that Complexism subsumes both, and analyzes the relevance
of evolutionary art practices in the context of the Complexism
movement.
The final part of the book comprises chapters
that focus on relatively unexplored areas of evolutionary art
and on the identification of future trends and open problems.
The sixteenth chapter, by Craig Neufeld, Brian J. Ross and William
Ralph, describes the evolution of artistic filters. The use
of multi-objective optimization techniques and of a bell curve
model of aesthetics, based on the empirical evaluation of artworks,
are some of the key contributions of this work, where a correlation
between aesthetics and the application of the paint operator
is shown. Gary R. Greenfield surveys co-evolutionary approaches
to evolutionary art, making a detailed description and analyzing
several instances of this type of approach. This analysis is
followed by a discussion of the challenges, difficulties and
opportunities posed by this type of approach. In the eighteenth
chapter, Penousal Machado, Juan Romero and Bill Manaris describe
a novel autonomous evolutionary art approach, where the competition
between an artificial critic and an evolutionary creator leads
to stylistic variation, presenting and analyzing the results
attained across iterations and in validation experiments. In
the closing chapter of the book, Jon McCormack looks into the
future, examining the challenges and possibilities that lie
ahead. He identifies and discusses several of the open problems
of the field from a research and artistic perspective, presenting
the background and motivation and discussing the theoretical
issues involved.
Finally, the DVD of the book comprises demonstration
programs, source code and valuable examples of images, music
and videos that complement the materials presented throughout
the chapters, allowing the reader to fully appreciate some of
the evolved works.
As previously stated, evolutionary art and
music research is reaching maturity, and part of this process
is the growing awareness of the various social, artistic and
scientific challenges the area faces.
The biggest social challenge for evolutionary
art and music lies in the development of projects or tools that
have a relevant social impact. Constructing tools that enhance
or promote the creativity of the user is probably the most obvious
way to address this goal. However, it is not sufficient ---
it is equally important to disseminate these tools and to improve
the public's awareness of their potential.
From an artistic perspective, the acceptance
of the evolutionary approach as a significant art practice is
probably the greatest challenge. To meet it, it is particularly
relevant to promote the participation of the artistic community
in biology-inspired endeavors, disseminate evolutionary projects
through the conventional art channels, and ensure their presence
in the commercial art circuit. Although some evolutionary art
practitioners and musicians have attained all of these objectives,
the challenge the area faces is ensuring that these exceptions
become the norm. The creation of art spaces devoted to evolutionary
art may play an important role in attaining it.
From a scientific standpoint, the development
of autonomous fitness assignment schemes that take into account
aesthetic criteria, the creation of systems that are able to
develop their own aesthetic concepts, the integration and interaction
of these systems with the environment, embodiment, and the definition
of new forms of human--machine interaction, are some of the
most relevant challenges.
Acknowledgments
This book would never
have become a reality without the enduring dedication of many.
We would like to express our gratitude towards all the authors,
who made this book possible, not only giving us such high quality
materials, but also offering their enthusiasm, perseverance,
patience and assistance. We would also like to thank Springer's
editorial staff, and especially Ronan Nugent, who provided encouragement
and support throughout the entire process. We also acknowledge
some of the early enthusiasts of this project, including Amílcar
Cardoso, Ernesto Costa, Alejandro Pazos, Antonino Santos, and
the EvoMUSART and Evo* communities. Finally, we are thankful
to Jorge Tavares, Santiago González and Eva Celeiro, who provided
valuable help in the revision and editing tasks; and to Erika
González, who designed the DVD.
A Coruña
Coimbra
August 2007
|
