“I see the awakening of consciousness as a series of spaced flashes, with the intervals between them gradually diminishing until bright blocks of perception are formed, affording memory a slippery hold.” Vladimir Nabokov
Pattern Recognition (2016)
Pattern Recognition is a performance for two dancers and eight robotic spotlights exploring themes of perception, memory, learning, and yearning — in both humans, and machines.
The piece is ultimately a meditation on how we project and breathe ‘life’, ‘intelligence’, ‘agency’, and even ‘consciousness’, onto animate objects, summoned by the ways they move and respond.
The robotic lights are not there merely to illuminate. They are ethereal sculptural performers in their own right. Their movements are not preprogrammed but arise generatively in real time, responding to dancers tracked by sensors, and driven by custom software built on artificial neural networks (artificial intelligence), running within a real-time, performative context.
Over the course of the 45-minute performance, the system learns from the dancers. At first, their responses are naive: they attend only to isolated aspects of the dancers, such as certain limbs, positions or orientations, and they merely mimic these features. Gradually, they begin to perceive how the dancers interact with each other, and from these observations they develop more complex, non-linear responses, creating an evolving choreography—a dance between human and machine.
While a human choreographer scripts the human dancers, the learning machine choreographer responds, deepens, and generatively transforms it, controlling the light dancer. The piece emerges as a third voice, a dance collaboratively written and performed, and unfolding generatively in real time, which neither human nor machine could have created on their own.
A collaboration with choreographer Alexander Whitley.
Choreography: Alexander Whitley
Digital Design and Concept: Memo Akten
Music: Scanner & Oliver Coates
Costume Design: Jean-Marc Puissant
Performance excerpts (5 minutes)
R&D Excerpts
Machine Learning R&D
Photos
Acknowledgements
Pattern Recognition is a Sadler’s Wells commission and DanceEast co-commission. We are very grateful to our co-producers, Hessisches Staatsballett / Staatstheater Darmstadt & Hessisches Staatstheater Wiesbaden, and MA Scène Nationale – Pays de Montbéliard and the generous support from Arts Council England, Panalux, PRS for Music Foundation, Trinity Laban Conservatoire of Music and Dance, South East Dance, Goldsmiths University of London, Jerwood Space, Middlesex University, Greenwich Dance Agency, Prettybird UK and Charles Glanville Dixon. Special thanks to Juliette Larthe.
Photos by Christopher Corless