PROgraMmable integrated photonic nEuromorphic and quanTum networks for High-speed imaging, communications and sEcUrity applicationS

Bio-inspired and quantum paradigms have risen as the two main competitors aiming to address the inefficiency of “conventional” computational schemes. Up to now these two paradigms utilize non-overlapping routes towards addressing this common objective; routes that are challenging in terms of scalability and integration readiness. PROMETHEUS’ proposition aims to shatter the boundaries between neuromorphic and quantum engineering and merge them to a common disruptive photonic integrated platform, able to merge each field’s merits. The platform will be based on a dense reconfigurable silicon waveguide mesh, powered by non-volatile BaTiO3 phase-shifters offering marginal power consumption and GHz tuning speed, during mesh reconfiguration. A key “upgrade” to this “synaptic” platform will be the co-integration of ultra-fast, multi-section lasers, acting as “neurons”. Sophisticated electro-optic packaging, will provide the basis to PROMETHES’ platform so as to evolve from the proof-of-concept limbo to a standalone neuromorphic/quantum hybrid integrated engine. This versatile chip will enable the realization of neuromorphic architectures such as large-scale spiking neural networks by exploiting signal multiplexing and GHz firing rate, photonic time-based convolutional networks, liquid state machines and on-chip training. PROMETHEUS will render feasible the exploration of disruptive concepts such as quantum reservoir computing. The platform will be utilized as an integrated quantum random generator and as a physical unclonble function, adding physical layer trustworthiness and novel security features. PROMETHEUS’ chips will be benchmarked, in terms of consumption, accuracy, speed and versatility, in a set of demanding applications that extend from ultra-fast image processing for flow cytometry and machine vision, signal equalization and authentication in high-speed optical links to quantum key generation and distribution.