Interior Images III
Computer installation with projection, 2002

Artificial Neural Networks technology belongs to significant branches of artificial intelligence research. It involves constructions, based on the model of biological nervous systems, which can be “trained” to execute certain special tasks via a kind of learning process. They function well when it comes to resolving tasks that require high capacity of generalization and low sensitivity to noise.

An Artificial Neural Network, trained to sense motion, processes the live image of the camera. The program produces the projected image on the basis of the output values provided by individual neurons. The interior changes of the artificial neurons, generally invisible to the human eye, are separated into several visible layers and projected into space.

With the artistic program I have followed for the last few years, I have two parallel aims in view. On the one hand, I intend to prepare models of the nervous system (in the form of computer programs) that are of high standard both in the technological and the scientific sense. These programs always attempt to simulate the functions of the biological nervous system -- a task that is rather difficult for traditional computer programs to perform. On the other hand, I “turn these models inside out”, i.e., I render their inner functioning visible in three dimensions: I make a sort of kinetic sculpture, whose movements of change are not controlled by coincidence, nor by a preset program, but by the continuous operation of an Artificial Neural Network.

Doing this I have the idea in mind that if we manage, however partially, to model the functioning of the human nervous system, we may take a further step in understanding existence and the act of creation.

Gábor Győrfi