Fatum (lat.) meaning: “In the eyes of the ancients, your fate was out of your hands – what happened was up to gods and demigods.”

Fatum is escape of our everyday reality. Escape from Covid-19, climate change, raise of populism and nationalism, trumpism etc. etc. . The aesthetics of Fatum are meta study into mechanism of learning progression with neural network models. Can audience experience the same thrill and attraction, what I’m sensing when seeing new generated video clip for the first time. Clips are fragments, glimpses of the trained AI model’s micro universum and materialistic essence, making unvisible visible.

Project started in May 2020 with interest towards predicting next frame of video material. I was inspired by examples seen in social media, specially Mario Klingemann’s Twitter post of fireworks created with pix2pix AI algorithm. Specially fashinated of sense with time and space.

With Fatum I am experimenting with different kind of media material; still images, short video clips, kinetic popup sculptures, 3d animations. Recently the focus has been in 3d modelling, -animations and iterating, recycling textures created by the modified pix2pix algorithm. One aspect has also been to explore glitches and imperfectness of the AI algorithm. And study prediction of the movement in kinetic sculptures and later with 3d animations.

In late fall 2020 I started collaboration with media artist Hannu Töyrylä, who programmed some new features to the original pix2pix code and also mentored to me image making techniques with GAN convolution layers. These improvements with code, made possible to do microsteps with gan epochs and also to mix convolution layers of trained models, and generic improvement of pix2pix’s next frame prediction algorithm. I also started to explore more how to mix various media materials in same dataset, implementing techniques of VJiing, diving deeper into evolution of the GAN image .

Fatum is predicting the future, envisioning the next frame. It has relation to time in various aspects. Evolution and iteration of the images, switching roles of input and output. Joyful play with epochs, frames and image sequences.

Video links are for preview only.

Fatum – Tuuli

AI predicting continuation of a video material. Length 45 seconds.

Technique: Photograph, AI algorithm.

https://vimeo.com/500884624/d9b33f3cc5c

Fatum – Marokkolainen

AI predicting continuation of a video material. Length 60 seconds.

Technique: Water colours, AI algorithm.

https://vimeo.com/500879472/6c609d449c

Fatum – Flowers, noise

AI predicting continuation of a video material. Length 60 seconds.

Technique: Photograph, AI algorithm.

https://vimeo.com/427456565/8587ae8987

Fatum – Harmaa

AI predicting continuation of a video material. Length 60 seconds.

Technique: Generative image, AI algorithm.

https://vimeo.com/472708008/2e8eee1270

Fatum – Rihmasto (Mycelium)

AI predicting continuation of a video material. Length 90 seconds.

Technique: Photograph, AI algorithm.

https://vimeo.com/453260570/9efbc6c130