In this project we generate a predictive biography of Clara Boj and Diego Díaz for the year 2050 made from all our digital activity (locations, digital conversations, photos, videos, etc.) collected during 2017. Year in which we hacked our mobile phones with a spyware software that captured all this information to create the Data Biography work consisting of 365 printed books (one for each day of the year with a total of more than 48 million records) and a 24-hour film story. In this new project we use this data to train different deep neural networks (RNN, CNN and others) in order to generate our predictable activity during the year 2050 and we have reprinted all this data to create another set of 365 books that make up the work Machine Biography; a fictional and predictive biography created by artificial intelligence where the limits of what is true and what is false are blurred, and the predictive capacity of algorithms and their social impact are questioned.

References/Credits:

– Time Series Forecasting, Long Short Time Memory ( CNN-LSTM ) multivariable and multisteps,

– GPT-2 Spanish Flax community, https://huggingface.co/flax-community/gpt-2-spanish with FineTuning by Mathias Gatti example https://www.cceba.org.ar/medialab/pueden-escribir-poesia-las-computadoras

– Dall-E Mini https://huggingface.co/spaces/dalle-mini/dalle-mini

– CUT CycleGAN2017, Contrastive Learning for Unpaired Image-to-Image Translation, Zhu, Jun-Yan and Park, Taesung and Isola, Phillip and Efros, Alexei A, IEEE International Conference on Computer Vision (ICCV), 2017, https://arxiv.org/abs/1703.10593

– Only a Matter of Style: Age Tranformation Using a Style-Based Regression Model by Yuval Alaluf, Or Patashnik, Daniel Cohen-Or, ACM Trans. Graph. 2021 https://arxiv.org/abs/2102.02754

– Fast-SRGAN, Hasnain Raza https://github.com/HasnainRaz/Fast-SRGAN