When Emotions Become Form (2019) is an hommage to Harald Szeemann’s exhibition “When Attitudes Become Form” whose message signifies not only in its year 1969 but until today. In the exhibition, Szeemann highlighted the importance of the process of conceptualizing the theory more than creating the artwork itself, which accentuated the role of curators as much of that of the artist in the realm of modern art. We also share the same notion by expanding such role to machines, proving that “(human) emotions” can be expressed in “AI-form”. Our artwork presupposes color and plane geometry as fundamental elements of visual arts.
By training via GANs (Generative Adversarial Networks), the AI analyses various images labeled with human emotions and generates obscure shapes and colors in layers through installation to present the ability of machine learning in interpreting abstract images such like such as the Korean abstract painting Dansaekhwa: We first consider the norms for positive and negative emotion-laden words and their color associations. And we used lines or circles on a virtual canvas and achieved different results for marks on a page under various conditions of emotions. Particularly, “Amusement” was found to be the most consistent with “Dialog” by Lee Ufan in 2011.
https://whenemotionsbecomeform.github.io/
