sergio albiac generates iterative dutch oil paintings through coded algorithm

sergio albiac is known for his use of computer coded algorithms as a tool for the generation of abstract portraiture and visual art. his work has included the employment of hubble telescope imagery in ‘stardust’ as well as data collected in real-time based on individual personalities in ‘I am portraits.’ while albiac’s work is born of contemporary digital processes of machine-learning, his next endeavor is a turn toward an historical golden age of oil painting, as the artist develops a generative algorithmic script on dutch oil masterpieces.

in this latest series entitled ‘you have learnt nothing,’ albiac proposes: ‘emerging from the shadow of time, they came to praise our scientific and technological prowess. queens and noblemen then and state-of-the-art hallucinations now, these people seem to smile at our contemporary propensity to confuse knowledge with wisdom. their faces are as grotesque as our ability to repeat mistakes.’

the iterative series by sergio albiac is the result of a custom ‘generative adversarial network’ (GAN). a GAN is a process of data input and machine learning involving two artificial intelligence algorithms: a ‘generator,’ which aims to precisely replicate a pre-existing image, and a ‘discriminator,’ which determines which input image is authentic. the discriminator receives two batches of input — the original, ‘real’ image together with that which is produced by the generator. through an iterative process the generator learns to accurately replicate the original input until the discriminator can no longer detect its inauthenticity.

text