previous up

Seminar: Between Images (2020)
Multilingual Typography
Jochem Cian

As a graphic designer I’m curious about how deep learning can be used in design and visual communication. After some experiments with different deep learning methods I asked myself, where I can see the potentials in machine learning processes.

Through deep learning, connections between different parameters can be made and lead to new innovations.

This finding lead me to the idea of merging different scripts like Latin, Cyrillic, Arabic and Hiragana (the japanese script system) to generate a new multilingual writing system with the help of AI. Therefore I trained StyleGAN with single letters of these different writing systems. Since StyleGAN is a texture-based network, I got very abstract letterforms as output.

As a continuation of my experiments I consciously decided against a concrete implementation of this typeface. I wanted to use the output to show the extent to which AI can be used in typography.

Therefore I created a moving poster for a fictive typography conference on the topic ›Multilingual Typography‹. The poster is based on moving, multilingual letterforms, which came out of my experiments. The blurring of the letters should inspire new ideas in type design.