Reinforcement Learning Agent Design and Virtual Reality Experience, 2019
Eccentric Nature explores the state of being connected with other non-human entities. A multi-level and multisensory immersive VR worlds is created to turn the audience into part of the emergent network. The agents in VR space are designed with genetic algorithm and reinforcement learning to simulate behaviours of various artificial and biological species, such as micro-organisms, forests, and AIs, and allow users to embody themselves inside those intelligent agents. The virtual worlds interpolate and blend between each other, moving and shifting from one modality to another, and making new possibilities of multi-sensory experiences.
Eccentricity + Nature
“Eccentricity” describes a state of movement off from the central axis, and implies the tilted, imperfect and sometimes turbulent geo-political climate that is constantly drawing unpredictable boundaries in our life. Its broken form "eccentri-city" denotes a collective state of being eccentric in the public domain, a city of eccentric culture that is deviating from the centripetal governing force, and a speculative future where creative energy is transformed into a counter-force to push the boundary and define new territories.
“Nature”, on the other hand, provides a systemic and ecological New Materialist perspective, questioning and examining the intra-activity among different biological species, artificial objects, emergent machines and humans. For the sake of better symbiosis, instead of leveraging and exploiting the nature with power, we link ourselves with the entities beyond our perception, converse and empathize with them in a new sensational network.