With this work, the artist seeks the representation of her childhood dreams. Through the process of training custom generative adversarial networks on a curated dataset combining personal photography and artificial imagery the meanings get obstructed and forms are lost. To this end, the artist explores the early stages of network training, where the main components of data distribution are imitated but not fully understood and captured. The liquidity of the abstraction alludes to how AI serves as a form of modern watercolor - one can impact but not fully control its flow.

Part of the Generative Abstraction show by Verse & EXPANDED.ART.