Information processing by a dissipative neuromorphic material

Artificial intelligence continues to influence all aspects of human existence. The energy consumption associated with digital computing is approaching its limits. Present computing technology employs the von Neumann architecture, characterized by distinct physical separation of logic and memory components. The frequent transmission of data between logic and memory demands substantial energy consumption. In this context, neuromorphic computing is regarded as a viable solution to fulfill future computational energy demands, as it integrates memory with logic. Neuromorphic computing draws inspiration from the human brain for information processing. This thesis demonstrates that dissipative self-assembly processes information similar to the actin remodeling in neurons. A dissipative material that self-assembles on a chip in response to electrical input signals has been shown to imitate the structural plasticity observed in human neurons while encoding information. Material utilizing photonic input to emulate synaptic plasticity has also been developed.