Proposal’s context, positioning and objectives
Our general goal is realizing photonic quantum reservoir computing (QRC) operating as a fully-functional next generation neural network (NN) computing system exploiting quantum physics. Leveraging unique photonic advantages enables the scalable parallelism essential for network-based computing, while quantum brings exponential dimensionality of a Hilbert space to NNs. PhotonicQRC strongly strengthens France’s and Germany’s position on the global novel quantum computing landscape by (1) developing quantum NN architectures optimized for photonic hardware, (2) establishing fundamental understanding of QRC prospects and limitations based on extensive theoretical investigations considering imperfections of the photonic hardware platforms, (3) extending quantum computing (QC) hardware by demonstrating the uniquely hardware-friendliness of reservoir computing (RC) in a quantum context, and (4) introducing benchmark evaluations to demonstrate quantum RC advantage. This interdisciplinary project’s impact will go far beyond the QC and NN research communities.
(a) Quantum nodes are randomly coupled, input is projected onto the quantum network, readout weights create the final result. (b) Parametric down conversion or quantum dots in micropillar cavities create quantum nodes. (c) Reservoir connections are created via complex transparent media or 3D photonic waveguides.