CLIP & VQGAN–generated memo and interpretation game.
Can you find pairs?
And can you tell, which text prompts belong to which images?
Following the last year's Strange Attraction StyleGAN-generated pexeso, this 2nd edition of pexeso games is based on a new trend in AI-driven image synthesis: text-to-image (CLIP&VQGAN) obsession.
This pexeso is a human-AI collaboration: text prompts created by humans in the form of positive affirmations and unique visualizations of these utopian concepts by artificial intelligence. The year 2021 demonstrates the need to imagine alternative futures. We believe that the first step for any idea to manifest is to first visualize it.
There are two extra features, that make this edition more advanced:
1. The matching pairs are not identical. They are both the result of the same text prompt, but a different iteration of the generation process. This way you can try out the AI visualization process at different stages and train your own neural network for pattern recognition.
2. Once you have completed the memory game, you can try the interpretation game - guess which text prompts belong to which images. (The used text prompts are part of the package).
64 pieces (32 pairs), 51mm x 51mm, 250g/m2 paper
Comes in 2 folded sheets (fits A4 envelope).
Lenka Hamosova is a designer, researcher, and visual artist based in Prague, Czech Republic. She is researching synthetic media and their use in audiovisual production within the doctoral study at Film and TV School of Academy of Performing Arts in Prague - FAMU. As the initiator of the educational project Collective Vision of Synthetic Reality (http://syntheticreality.hamosova.com), she spreads awareness of available AI models and opens a discussion on the future use of artificial intelligence through interdisciplinary workshops.