Proteus 3.0 - Interacting with the cloud

SHORT PAPER. This paper presents the Proteus 3.0 project as an interactive and generative video installation, which tries to remain in conversation with its audience and learns from its engagement to produce new and potentially meaningful sequences of images. This particular version of a series is focused on digitized material as ferrofluid and its digital surrogate generated with a reaction-diffusion algorithm, projected onto a room-scale oval oculus allowing for a collective/immersive interaction. By employing state-of-the-art custom reinforcement learning models coupled with human intelligence in a symbiotic fashion, it aims to make the interaction more meaningful by understanding better human behaviour during an interaction. Additionally, it intends to maintain the interaction as conversations, rather than playful attractions.