|
|
Liquid Perceptron (2000)
Hans H. Diebner
Sven Sahle
Work Description
The brain is a network that consists of 10 billion neurons. On the one hand, it works as a stimulator and, on the other side, as a simulator. Neurons can be excited through external stimuli and thereby led to an oscillatory state. Via axons and synaptic connections these stimuli are transmitted to adjacent neurons. This leads to a global pattern formation of brain activity.
The brain remains in activity even if there is no external stimulus, which means that it simulates, for example, when we are dreaming. However the stimulator and simulator cannot be strictly treated separately. One may think of the famous deprivation tank experiments where the test persons who are fully de-coupled from external stimuli very quickly start to halluzinate.
The installation Liquid Perceptron is a simulation of a neuronal network that is excitable through the external reality. Specifically, the simulation of the neuronal network is excitable through the movement of the spectators. Comparable with the eye, a camera recognizes the movement of objects in space and transmits the video signal to the neuronal network. The open loop state of the brain shows wave-like fluid patterns. After the network has been coupled to the external world an image of the reality becomes emergent - superimposed by the "eigen dynamics" of the Liquid Perceptron. The activity of the neurons is color-coded. Bright areas are particularly active. One recognizes that local excitations lead to the formation of coherant patterns, much like observed in brain activity visulizations. The activity fronts spread of the whole network. The information corresponding to ’’perception’’ is represented in the global activity patterns rather than in each single neuron.
It has first been recognized by Alan Turing that small local instabilities, perturbations or excitations, respectively, within a part of a multi-component coupled system lead to a spread of coherent global patterns throughout the total system. In his very important paper on the chemical bases of morphogenesis in 1954 he already addressed the still open question where the information is located that tells the ovum where the head and the feet, respectively, of an embryo have to be placed.
A further example of the same category of dynamical systems are the formations of structures in furs of animals where in the same manner so-called "Turing-instabilities" lead to global pattern formations in the total system. Compared with morphogenesis the time scale of the pattern forming processes in the brain is much smaller and, additionally, these processes are only temporarilly stable. The brain is almost always in a transient state. Nevertheless, from a dynamical point of view morphogenetic systems and brains can be regarded as closely releated. The series of three images shows subsequent snapshots of a morphogenetic pattern formation process within a two-dimensional array. One sees that the pattern (so-called "Turing spots") gradually emerges and eventually becomes stationary.
Liquid Perceptron allows both the researcher as well as spectators to perform experiments in order to achieve understanding of how complex systems work. The interaction of spectators with Liquid Perceptron is itself a valuable source to understand cognitive functions. [H.H.D.]
Keywords
brain dynamics, artificial intelligence, pattern formation, synergetics, self-organisation
|
|
Aspects of Liquid Perceptron
Hans Diebner: Liquid Perceptron. Installation View (2000)
Liquid Perceptron Documentation
Observe a video documentation of the Liquid Perceptron → here
(c) Hans Diebner 2004
Hans Diebner: Liquid Perceptron. Stage Background for the Opera of ’’Einstein on the Beach” from Philip Glass and Robert Wilson.
August 5th 2005 in Berlin (D).
Production: Berthold Schneider
Curator: Ralf Hartmann.
additional information
Hans Diebner: Liquid Perceptron. Installation View (2000)
Additional Links
Works
Institutions
Performative Science
Basic Research
INM Institut für neue Medien Frankfurt
|
