Looking Outwards 02 – Learning to see

“Learning to See” is an ongoing work by Memo Akten. The artist describes this work as using machine learning algorithms to reflect the way that human perception reconstructs the world around us based on our own beliefs and experiences rather than “accurately” representing the world around us. I think this is a really interesting idea for a project and that the “vision” of the machine learning algorithms makes for a powerful metaphor for the aritst to make his point. I also find the real time fluid interation between the image generated by the algorithm and the input objects inspiring. It is a really interesting hybridization of intuitive manual gestures and an algorithmic art making process. I don’t know a whole lot about Generative Adversarial Networks (GAN), which is the type of algorithm being used in this piece. However, I do know that it has to be trained on a database of imagery which determines the sort of images that it creates. We can also asume that there is somewhere within the process a video feed is being fed into the algorithm to allow for the output to be manipulated in real time.

Learning to See (2017-)