Symbolic Grounding is a kinetic light and sound installation that runs at its core by a network of 100 artificial analogue neurons. In this network function, temporal structures, ordered sequences and coordination patterns emerge from the interaction of the artificial neurons and the connectivity of the network. The installation renders these temporal structures and functions tangible and allows the visitor to experience the complexity of interactions happening in such a network.
Our approach is guided by systems theory, one the one hand theoretical models that explain inter-limb coordination, specifically the coordination dynamics of finger tapping and on the other hand by models from theoretical neuroscience.
Similar to the technique of using analog computers for solving complex equations , we use analogue electronics to run the equations for describing the temporal evolution of models of neural populations as proposed by Shun-Ichi Amari.
Unlike using the algorithmic software libraries for executing the feed-forward neural networks as they are provided by Google we are interested in non-algorithmic approaches to computation. In the systems we develop there is no central clock that steps through a set of instructions, no reading or writing of symbols as in the Turing machine, but a continuous flow of time and electrons. Our tools are the methods and equations as they have been developed for describing circular systems that feed back onto themselves by dynamics systems theory and earlier by cybernetics. We use these tools in a constructive and experimental approach to produce temporal patterns that may seem algorithmic, but are not.
A neural Oscillator based on the Shun-Ichi Amari’s work by Christian Faubel and Wolfgang Spahn.
Its an analog electronic implementation of the model described in “Dynamics of pattern formation in lateral-inhibition type neural fields” by Shun-Ichi Amari.
Simulation of two coupled neurol oscillators by Christian Faubel.