Creative Machine | Program
A major exhibition exploring the twilight world of human/ma- chine creativity, including installations, video and computer art, artificial intelligence, robotics and apps by leading artists from Goldsmiths and international artists by invitation.
archive,category,category-program,category-125,ajax_updown_fade,page_not_loaded,,select-theme-ver-1.2,wpb-js-composer js-comp-ver-4.3.4,vc_responsive

The Performance

Lillevan & Parkinson

This is the first time Lillevan and Adam have performed together. They will improvise with a set of prepared materials.

Lillevan’s practice investigates non-narrative facets of film, focusing on the musicality of the imagery, thus defining the moving imagery as an instrument in its own right as opposed to accompanying music. Intensity and texture are more important than narrative and figure. The relationships between the image’s elements and the viewing eye, between the eye, the mind and the soul are explored. The world of media archaeology is of major interest, while questioning viewing habits and manipulative image-creation. Lillevan recontextualises, combines and politicises existing film images and fragments. The images are a communicative medium interacting with the music. The selection of the images can either support the sound, or work against it, the aim being to achieve a dialogue.

Adam will be performing using 3 Beagleboards (single board computers): bare circuit boards which are essentially computers without screens, housing or keyboards, running granular synthesisers and algorithmic, unpredictable sequencers.

A large part of his musical practice involves recontextualising and reappropriating samples, taking sounds from easy listening, hardcore trance and 80s pop and forcing them into unfamiliar combinations, trying to reveal their strange and hidden potentials when placed in new contexts. For this performance he will be working with sine waves and sounds synthesised on an MFB drum machine, sonically occupying the space with a limited sound palette of skeletal techno.

Light sensors mounted on the Beagleboards create unstable connections between Lillevan’s images, reduced to vectors of intensity, and the soundscape, allowing both performers to interact with the sound and find areas of expressivity and conflict.

The performance will take place at the Private View for the Creative Machine exhibition on Thursday 6th November in the SONICS surround audio visual installation in the main exhibition space.


Tantalum Memorial

Harwood, Wright & Yokokoji

Tantalum Memorial was produced by Harwood, Wright and Yokokoji between 2006 and 2009. The work is a telephony-based memorial to the people who died as a result of the ‘coltan wars’ in the Congo from 1998 to the present. Coltan ore is mined for the metal tantalum – an essential component of mobile phones that is now more valuable than gold. The work is constructed from redundant electromechanical Strowger switches – the basis of the previous generation of telephone exchanges.

These switches are reanimated by tracking the phone calls from Telephone Trottoire – a social telephony network also designed by the artists for the Congolese radio programme Nostalgie Ya Mboka in London. Their precisely poised movements and sounds create a sculptural presence for this otherwise intangible network of circulating conversations and weave together the ambiguities of globalization, transnational migration and the impact of our addiction to constant communication.


Paul-IX le vaniteux

Patrick Tresset

Paul-IX le vaniteux, passes time by drawing a still life from observation. The ensemble of objects depicted seems reminiscent of a Vanitas of the XVIth century; a type of motif traditionally depicting objects that symbolise different aspects of the futility of human earthly pursuits.

The irony of an artificial agent commenting upon human behaviours, aspirations and mortality is counterbalanced by the knowledge that, just as the Nexus-6 in P. K. Dick’s Do androids dream of electric sheep the Paul series of robots, if not maintained, have a short life expectancy. Anyway, what is the point for a robot to dedicate its time to criticising human existence rather than be a useful machine as expected of it?

The Paul series of robots are artificial agents obsessively focused on the drawing practice. Paul predecessors were originally developed to palliate a debilitating painter’s block and as such can be seen as creative prosthetics or behavioral self-portraits.

Even if the way the Pauls draw is based on Tresset’s technique, their style is not a pastiche but rather an interpretation influenced by the robots’ characteristics. The software driving the Pauls is based in part on technologies developed by Tresset in the context of AIkon-II, a research project hosted within Goldsmiths College’s computing department.
The AIkon-II project investigated the observational sketching activity through computational modeling and robotics, and was co-directed with Frederic Fol Leymarie.


Space Flower series

Naoko Tosa

Naoko Tosa’s new series of works pays homage to Rimpa, one of the major historical schools of Japanese painting that was founded in Kyoto in the 17 th century.

Famed for its usage of luxurious golden colour, one of its key exponents was the artist Korin Ogata, whose paintings of Wind God and Thunder God have become representative of the style. The fragmentation of flowers by the elements of wind and thunder serves an allegory for the fading of old customs and the beckoning of a new future.

Oiran alludes to the eponymous courtesans popular during the 18th and 19th century; the roses, as though performers in a Kabuki piece, exude a gallant beauty while their fragility evokes a dreamlike presence, at once ephemeral and elusive. A flower blooming in space in Space Flower symbolizes life and refers to one’s individual self. Perhaps the act of maturing requires one to reject one’s past, an act that inflicts hurt upon one’s self, but the individual more often than not comes out stronger. Space Jungle depicts a jungle on a planet far away from ours.

Abundant water and flora abound on the surface while an array of minerals and plants inhabit the underground, navigating the chaos of zero gravity in obliterative spurts of dance. Drawing inspiration from Ogata’s painting Red and White Plum Blossoms, a national treasure of Japan, Moon Flower considers how red and white plum blossoms would appear on the Moon, presenting a continuous kaleidoscope of exploding moon flowers. The deified characters of Ogata’s Wind God and Thunder God take centre stage with Tosa’s reimaginings. Wind God shows a figure approaching and unfolding onto the scene before departing while Thunder God invokes the Japanese thunder god Susanoo, who remains likeable in spite of his selfish and mischievous personality.


Feedback Variations

Peter Todd

This work is part of an ongoing technical and artistic exploration of feedback-based processes. It represents a refinement of some of these techniques into a form which is minimalistic in its algorithmic structure while – like an early Steve Reich phase piece – beguilingly complex in its perceptual effect.

In contrast to the pervasive feature-creep that can infect software development, much of the process behind this work is concerned with the removal of elements non-essential to the pure expression of the underlying phenomena. As works in their own right, these stand coherent and fully-formed. The starkness of their largely black and white appearance is not a technical limitation, but a facet of the desire to retain clarity as well as a reflection of the underlying simplicity.

The techniques used need not necessarily have these austere limitations and indeed can – in different variations, not presented here – incorporate rich palettes of colour. Other uses may also complement more complex environments in peculiarly satisfying ways.

In the case of both of the variations displayed here, the process involves video feedback of the form where the results of the previous frame’s rendering are used as visual input to each new frame. This colour value is inverted and rendered onto some simple geometry which remains static while the viewer’s manipulation of the ‘camera’ through which it is viewed changes. The use of colour inversion helps to ensure that the results do not tend to ‘blow out’ or decay excessively as patterns converge or diverge, allowing wide ranging exploration of the space of possible patterns without too often being stranded in featureless desserts.

Also crucial to the design of these variations is the use of solid ‘frames’ of pure black (or black with a white stripe), which serve to provide the system with a consistent input signal encouraging the formation of stable patterns.

In Fractaleid (for iPad, with other platforms pending release at the time of writing), the geometry on which the feedback is rendered is a kaleidoscopic plane. The viewer manipulates a 2d camera with pan, pinch and rotate gestures on the touchscreen corresponding to movement, zooming and rotation. A black rectangular frame on the outside of the rendered image provides
a stable graphical input, while subtly coloured dots appear under users’ fingers; sometimes changing the entire appearance of the image as they are incorporated into the feedback.

The variation Cardboard Box Recursion (Android & Cardboard) explores the synthesis of a unique and compelling virtual reality environment distinct from resource intensive graphics techniques that attempt to produce a realistic, representational view of the world. It uses a 3d environment consisting of a cube, with the viewer placed in the center. The camera through which the feedback loop occurs uses a damped version of the viewer’s head movements, opening up an expressive dynamic interaction with the chaotic geometries as they unfurl into endless tunnels and spiralling voids. The ‘frame’ consists of a black and white stripe along each of the edges of the cube.


Different Ways to Infinity

D.W.I gathers experiments as fictional strategies to generate infinity: the complexity of chaos, the space filling capacity of the rhombic dodecahedron, the replication of motorized mechanical systems as clones, or a machine’s effort to raise a pendulum against gravity to find perpetual equilibrium.

The ancient alliance has been destroyed; man knows at last that he is alone in the universe’s indifferent immensity out of which he emerged only by chance. – Jacques Monod

In the late 70’s, chaos theory challenged traditional science’s mechanistic vision of the universe (La nouvelle alliance 1979). Reality became an open system based in change, disorder and entropy.

Following the Prigoginian paradigm, DWI Chaos and DWI Clones, focuses on those aspects of reality such as disorder, instability, disequilibrium and non-linear relationships. The installation confronts the viewer to the anatomy of nature as an open system: entropy, the metaquantification of complexity, gravity, the control of chaos…

DWI: Clones

Clones are represented as identical mechanical machines (inverted pendulums), that using human behaviors, fuzzy logic, drift to find endlessness equilibrium.

Clones uses two identical inverted pendulums: based in a motorized mechanical system, the pendulums are mounted with the pivot point on a cart that moves horizontally. This movement makes the pendulum swing, and once the pendulum rotates to its inverted vertical point, an algorithm takes control of the movement of the cart and tries to find equilibrium against gravity. The artist explores the expressiveness of the system, using human size heavy pendulums, capable of generating enough inertia to generate self-rotation. He also tries to humanize the system in a game of failures and successes.


Iconographies #018 – 
Venus & Adonis’ after Rubens

Iconographies is an ongoing project focusing on the analysis of renaissance and baroque paintings via computational methods.

Religious and mythological scenes are transformed into complex abstract formations.
By removing iconographic narratives, the paintings lose their original context to become new objects of contemplation.

Iconographies #018 is a large-scale print inspired by Venus & Adonis (1610), a Peter Paul Rubens painting – Museum Kunstpalast, Dusselfdorf, Germany.


Structure in Flux

Vesna Petresin

Investigating flow, rhythm and growth, the video piece was generated during a live performance Structure in Flux (Rubedo, 2010). Here, 3D structure is being shaped, in real-time, in response to frequencies, modulations and rhythmical patterns of the sound piece composed and performed during the performance.

3D structures rendered as fluid light react to the acoustic parameters of the music; this creates a constantly-evolving visual experience of the installation.

The outcome is alike a moiré pattern – a figural effect produces by superposing fields of sonic loops and light patterns, with complex behaviours and experiences resulting from a combination of regular, repetitive elements.

This transmediation of sonic to visual organisation makes it possible to create an ecology of sound and light – a system based on its capacity for variation in space and time.

The result is an evolving landscape of light and sound that allows the viewers to discover their own sense of order, and engage with emotions arising from the experience.

Credits for “Structure in Flux”: Rubedo (Vesna Petresin & Laurent-Paul Robert), 2010.


Intentionality in Silico (Sold Christ)

Parashkev Nachev

A pervasive source of conceptual anxiety is the notion of intentional in-existence: that something can exist in a mind yet nowhere else. For the reductive theorist it seems an attractive hook on which to hang the distinctiveness of the mental, for it seems both fundamental and unparalleled in the non-human world. Though we had little use for the idea before the scholastics brought it back to defective life, a conceptual attack on it is too deflationary for the current intellectual market. So here instead I turn to empirical, guerilla warfare, by creating a machine that confabulates images that exist nowhere but in its “mind” yet have the reality of that most human of imagined unreals: the inexistent face. Employing deep-learning techniques I have created a new kind of machine, a facievore, that consumes human faces it can find on the internet and extracts canonical, archetypal representations, automatically shaped by “imhomogeneities” in the population, pictorial or categorial. It consumes both face surrogates (photographs, masks) and face representations (painting, sculpture), sometimes confabulating from the real, and sometimes from the human-imagined. Though seemingly part accidental, part reality-driven, its complexity pushes it into territory where the distinction between the stochastic and the deterministic becomes opaque.

Aside from smiling at the implausible reductiveness of dominant ideas in the philosophy of mind, this machine has another, positive aspect.

For it draws on a truth to which machine-learning more potentently than any other set of ideas will awake us: that the domains of the human and the physical-biological are one, must be one, and so to understand the biological we must humanise it.

This work supports a Wellcome Trust & Department of Health funded translational research project to develop a clinical system for detecting anomalies in brain scans with the aid of machine-learning (HICF-R9-501).


Niche Constructions

Jon McCormack

Niche construction is a biological process whereby organisms modify the conditions and resources of their environment to create heritable niches for themselves and their offspring. For example, Beavers build dams which influence river flow and lake formation, creating a niche that changes plant composition and decomposition dynamics in the immediate area of the dam.

In the standard view of evolution, the gene is the main unit of selection in evolving populations of phenotypes. Species develop adaptations to their environment and the responsible alleles proliferate in the gene pool. From a niche construction perspective, organisms modify their environment – and possibly that of other species – creating a heritable environment for offspring. Advocates of niche construction argue that this forms an important feedback process in the natural evolutionary process.

This work is part of a series of experiments where we have applied biologically inspired processes to the development of creative generative systems.

In Niche Constructions, line-drawing agents move over a virtual canvas, leaving a trail of ink as they move. While drawing, they might reproduce, giving birth to new lines with similar drawing behaviour. If an agent intersects with an existing line, it dies. Eventually, all species become extinct; the drawing space fills up, all the agents die off, and the drawing is complete.

Each agent’s genetic structure contains a special gene that represents the optimum image density the each individual and its descendants. The value of this optimum is genetically determined, but can vary between different species of agent.

As the agent moves around the canvas, the density of ink is measured in a small area surrounding the agent’s current position. This density measure is used to determine how suited the environment is to the agent’s density preference. The closer the match, the more successful the agent is in terms of longevity and reproduction.

Over the life of the drawing, the agents modify the environment to suit their niche preference. Additionally, through evolution, the genetically determined density preference adapts as the drawing fills with lines and becomes more dense.

The niche construction drawings show much greater variation in density and drawing style than the drawings made by agents without this mechanism. Niche construction introduces more complex behavioural dynamics into the drawing process using a relatively simple mechanism. ‘Founder’ agents often draw large, closed boundaries to protect their low-density children from being invaded by high-density loving invaders, for example.

We often think of creativity as a complex cognitive function that is most developed in humans. Niche Constructions shows how the bottom-up interaction of many simple, low-level processes can result in a system that exhibits creative behaviour.



Manu Luksch

The sci-fi film FACELESS plays in an eerily familiar city, where the reformed RealTime calendar has dispensed with the past and the future, freeing citizens from guilt and regret, anxiety and fear. Without memory or anticipation, faces have become vestigial – the population is literally faceless. Unimaginable happiness abounds – until a woman recovers her face…

The film was made under the constraints of Luksch’s Manifesto for CCTV Filmmakers – images are obtained from existing CCTV systems by the director/protagonist exercising her rights as a ‘surveilled person’ under data protection legislation. To comply with privacy legislation, CCTV operators are obliged to render other people in the recordings unidentifiable – typically by erasing their faces, hence the ‘faceless’ world depicted in the film. The scenario of FACELESS thus derives from the legal properties of CCTV images: a ‘legal readymade’.

There was no traditional shooting script: the plot evolved during the four-year long process of obtaining images. Scenes were planned in particular locations, but the CCTV recordings were not always obtainable, so the story had to be continually rewritten.

The medium, in the sense of ‘raw materials that are transformed into artwork’, is not adequately described as simply video or even captured light. More accurately, the medium comprises images that exist contingent on particular social and legal circumstances – essentially, images with a legal superstructure. Faceless interrogates the laws that govern the video surveillance of society and the codes of communication that articulate their operation, and in both its mode of coming into being and its plot, develops a specific critique.


Cellular Forms

Andy Lomas

Inspired by the work of Alan Turing, Ernst Haeckel and D’Arcy Wentworth Thompson, Cellular Forms uses a simplified biological model of morphogenesis to explore the generation of complex three dimensional structures.

Each form starts with a small initial ball of cells which is incrementally developed over time, adding iterative layers of complexity to the structure. The aim is to create forms emergently from the interactions between individual cells, exploring generic similarities between many different shapes in nature rather than emulating any particular organism. The process reveals universal archetypal forms that can come from growth-like processes rather than top-down externally engineered design.

Cell division is controlled by accumulated nutrient levels. When the level in a cell exceeds a given threshold the cell divides, and various parameters control how both the parent and daughter cells re-connect to their immediate neighbours. Rules can also be adjusted for how nutrient is created, such as by being randomly uniformly created by each cell, or by incident light rays creating nutrient in cells hit by photons. Nutrient can also be allowed to flow to adjacent cells. The simulation process is repeated over thousands of iterations and millions of particles, with typical final structures having over fifty million cells.

A number of internal forces affect the structures, including linear and torsion springs between connected cells. Additional forces repel cells that are in close proximity but are not directly connected. This creates tensions within the structures that induce them to change shape dynamically, with surfaces naturally folding into complex organic forms.

A wide set of variations arise from small changes to the rules governing the systems, with selection of forms based on aesthetic considerations rather than optimizing a conventional fitness function. All resultant motion as well as shape is genuinely emergent, since the simulation rules only dictate interactions between adjacent cells.

Two complementary rendering methods are applied to the simulation data to visualize the generated data and reveal different aspects of the forms.

The first shows the cells as sphere primitives illuminated with a diffuse light from all directions, revealing the three-dimensional shape of the forms through self-shadowing of the surface.

The second uses a density accumulation map, with each cell represented as a sphere of equal density. This digital emulation of an X-Ray reveals internal details of the structures that may not be apparent from the external surface.