Latent Belonging contemplates on memory’s fragmentation particularly of past home spaces and its transformation as our sense of belonging emerges. With a DCGAN trained on a set of 659 personal archival home images, the dataset, limited by the number of photos I could possibly gather, consists of concrete places and experiences that are forced into abstract imagery. The machine’s ambiguous reinterpretation of the dataset mirrors the frustration in making sense of our fragmented memory. Exploring the artistic possibilities of machine malfunction, the latent space of the machine constructs an unstable image in process (softimage), comparable to the transformative nature of place-making and the emergence of belonging.