EvoMUSART 2014: Third International Conference on Evolutionary and Biologically Inspired Music, Sound, Art and Design

Evolutionary Computation (EC) is a broad class of computational techniques inspired by nature, and in particular by Darwinian evolution. EC is used for search-and-optimization problems ranging from business and engineering to graphical art, as described in this section. It may be familiar to long-time readers, as Leonardo has previously published two special sections on evolutionary art, in Leonardo Vol. 35, No. 2, and Vol. 36, No. 1.

The basic principles are those of genetics. In EC, as in nature, there is a population of individuals, and each has some genetic material (a genotype) that affects the appearance or behavior of the individual. Among the individuals of the population, some are “better” than others and give rise to new individuals by mating or mutation. Others are discarded. Over time, the population improves. In nature, better refers to the ability to survive and reproduce in a harsh environment. In EC, if the problem is one of engineering or business, better may be defined in computational terms. In an artistic domain, better may be defined with reference to a theory of aesthetic value or it may be implemented by interactive selection of individuals by a user based purely on subjective preference.

The use of EC in artistic domains is called evolutionary art. It has been used as a tool in the creation of many types of art, including music and musical timbre, graphical art and animation, 3D design, architecture and poetry. Evolutionary art is appealing to artists for several reasons. It can be used for exploring the implications of an aesthetic theory or for generating unexpected ideas. It allows a novel workflow very different from the typical workflow of most artists. It can allow peripheral, low-value tasks to be automated, leaving the artist to focus on core issues.

The EvoMUSART conference was born as a workshop under the Evostar (Evo*) umbrella in 2003 with the aim of helping the consolidation of evolutionary art and music research. Its scope is evolutionary and spans bio-inspired music, sound, art and design. Here, bio-inspired refers to methods that are related to but slightly outside of EC, including particle swarm optimization, ant colony optimization and neural networks. After 12 editions, the event is now a full international conference and has included more than 400 papers.

These papers represent the latest work on three projects selected from those presented at EvoMUSART 2014, including the Best Paper Award winner and runner-up. They reflect some of the main topics of evolutionary art, both interactive and noninteractive. Eisenmann et al. report on a novel mode of interaction for 2D and 3D evolutionary art. Their system allows the artist to express preferences for particular parts of the evolving designs rather than for whole designs. It uses sensitivity analysis to perform the necessary “backward mapping” from parts of the design to the genotype. It also aims to reduce “visual load” on the artist. Machado et al. propose another new mode of interaction: Their system allows the artist to specify the goals for a noninteractive generative art system using artificial ants, interactively and with real-time feedback. Finally, McCormack describes an interesting encoding for simple animated creatures, which was employed in an interactive museum exhibit to give each visitor a personalized experience. He uses it as the jumping-off point for a discussion of what properties of search spaces are desirable for artistic uses.

—Juan Romero
Universidade da Coruña, Spain
Email: <jj@udc.es>

—James McDermott
University College Dublin, Ireland
Email: <jmmcd@jmmcd.net>

Contents

Jonathan Eisenmann, Matthew Lewis and Rick Parent: Spatiotemporal Ideation and Generation with Interactive Evolutionary Design	246
Penousal Machado, Tiago Martins, Hugo Amaro and Pedro H. Abreu: Beyond Interactive Evolution: Expressing Intentions through Fitness Functions	251
Jon McCormack: Codeform: A Balancing Act between Variation and Utility in Evolutionary Art	257