neuromorphic art

Art generated by neural networks, yet possessing human-like aesthetics

Many images look as if they were created by people, even though none of the neural networks had exposure to any human artwork

All images are constructed by the networks from scratch

See Overview for more details

Physical (non-virtual) exhibition

Neuronal visions

23^rd May — 30^th June 2024, Library, Faculty of Physics, Astronomy and Applied Computer Science, Jagiellonian University, Kraków, Poland.

Featured online exhibition

Interpretations

Tap or click anywhere to enter

Variations

The neural networks showcased in this gallery exhibit rather painterly aesthetics. The images created by a particular network usually have a fairly consistent style — the network's individual “artistic personality”. Yet, there is surprising diversity among the various neural networks.

Tap or click anywhere to enter sigma networks

Dreams

These images have a surreal or even dadaist quality. Even though the imagery may seem haphazard, one can identify recurring motives characteristic of the original objects. These images can be understood as a visual analog of symbolic representations, reminiscent of dreams.

Tap or click anywhere to enter gamma networks

Expressionist

Vibrant colors and shapes, deformations, energy. This brings to one's mind Expressionist art, hence the label for this collection. The individual neural networks are much less stylistically consistent than the ones exhibited in Variations.

Tap or click anywhere to enter tau networks

Abstract

Textures and intertwined shapes bordering between the organic and technological. Yet staying just beyond representing anything tangible. Thus abstract, but only up to a point.

Tap or click anywhere to enter kappa networks

Graphics

The neural networks of this variety create images, some of which are surprisingly reminiscent of various graphic techniques. Others seem akin to Art Brut. Yet, altogether they evoke a strikingly consistent visual feel.

Tap or click anywhere to enter omega networks