Memristive Reservoirs Learn to Learn

Memristive reservoirs draw inspiration from a novel class of neuromorphic The synaptic sites of nanowire networks are not directly accessible, hardware known as nanowire networks. These systems in contrast to random access memories (RAM) [4, 9, 17], where display emergent brain-like dynamics, with optimal performance each memory cell is addressable and programmable. The lack of controllability demonstrated at dynamical phase transitions. In these networks, is compensated for by the dynamic nature of nanowire a limited number of electrodes are available to modulate system networks, which is a key feature that enables them to adapt to dynamics, in contrast to the global controllability offered by neuromorphic evolving input signals. Nevertheless, it is worth investigating how hardware through random access memories. We demonstrate these neuromorphic systems can be optimized for information processing that the learn-to-learn framework can effectively address this tasks. For example, previous studies have shown that in a challenge in the context of optimization. Using the framework, we physical reservoir computing framework, nanowire networks can successfully identify the optimal hyperparameters for the reservoir.