In Hypothetical Reconstruction: Seeds, Kevin Abosch uses photography and a generative adversarial network (GAN) to create a corpus of natural and synthetic seeds (Swiss chard) that individually and collectively serve as a proxy for the purpose of distill ing emotional value. In a ritualistic back-and-forth process, Abosch uses machine learning (ML) and his own intuition to generate, discriminate against, classify and prune visual data. This feedback loop persists until the artist is convinced that a lost or forgotten truth has been “reconstructed.”

Full Size