Am Neural Network Architecture

by SONAMB

This image represents a speculative neural network architecture where the sounds we hear in our dreams are extracted to be used as sonic training data whilst we are sleeping. This model trained on our private inner dream-sonic worlds crosses boundaries of consciousness and control and is able to output new sonic auras and realities previously uncharted. SLEEPSTATES.NET is a browser based work where the user navigates to different states of algorithmic slumber, a sleep-cycle of audiovisual interactions that takeover their screen. Exploring themes of machine addiction and autonomy, sleep sanctity, the networked condition and how citizens are subjected to emotion extraction via platform capitalism, these dream transmissions journey through the techno- emotional states of slumber we experience between humans and machines late at night. Users experience these AV screen-takeovers and interact with the interface via checking their slumber status and machine addiction levels. Finally they are asked to join SONAMB in defending sleep by ‘transmitting their frequency’ and breathing into the machine. An audio reactive noise field appears where they contribute to a collective white noise sleep aid. This net-art piece accompanies SLEEPSTATES, the debut album from SONAMB, the new music project from Manchester sound artist Vicky Clarke and features sounds from the record across the platform.