Reconfigurable qubit states and quantum trajectories in a synthetic artificial neuron network with a process to direct information generation from co-integrated burst-mode spiking under non-Markovianity

Abstract

The research and development of hardware neuron technologies are accelerating at a very fast pace to provide for increased efficiency in performing artificial intelligence and autonomy functions beyond that possible with emulation on digital computers. Moreover, dedicated hardware for these biologically inspired functions creates capabilities not possible currently—especially regarding the integration of quantum information processing and advanced non-linear dynamical phenomenon necessary for bridging the gap between artificial and biological intelligence. A synthetic artificial neuron network functional in a regime where quantum information processes are co-integrated with spiking computation provides significant improvement in the capabilities of neuromorphic systems in performing artificial intelligence and autonomy tasks. This provides the ability to execute with the qubit coherence states and entanglement as well as in tandem to perform functions such as read-out and basic arithmetic with conventional spike-encoding. Ultimately, this enables the generation and computational processing of information packets with advanced capabilities and an increased level of security in their routing. We now use the dynamical pulse sequences generated by a memristive spiking neuron to drive synthetic neurons with built-in superconductor-ionic memories built in a lateral layout with integrated niobium metal electrodes as well as a gate terminal and an atomic layer deposited ionic barrier. The memories operate at very low voltage and with direct, and hysteretic Josephson tunneling and provide enhanced coherent properties enabling qubit behavior. We operated now specifically in the burst mode to drive the built-in reconfigurable qubit states and direct the resulting quantum trajectory. We analyze the new system with a Hamiltonian that considers an integrated rotational dependence, dependent on the unique co-integrated bursting mode spiking—and where the total above threshold spike-count is adjustable with variation of the level of coupling between the neurons. We then examined the impact of key parameters with a longer-term non-Markovian quantum memory and finally explored a process and algorithm for the generation of information packets with a coupled and entangled set of these artificial neuron-qubits that provides for a quantum process to define the level of regularity or awareness of the information packets. These results, therefore, enable quantum neural networks where qubit/quantum memory states and the associated quantum trajectories are now available for conducting advanced computational algorithms in conjunction with the information processing capabilities of general spiking neurons.