Direct implementation of a perceptron in superconducting circuit quantum hardware

Abstract

The utility of classical neural networks as universal approximators suggests that their quantum analogues could play an important role in quantum generalizations of machine-learning methods. Inspired by the proposal in [1], we demonstrate a superconducting qubit implementation of an adiabatic controlled gate, which generalizes the action of a classical perceptron as the basic building block of a quantum neural network. We show full control over the steepness of the perceptron activation function, the input weight and the bias by tuning the adiabatic gate length, the coupling between the qubits and the frequency of the applied drive, respectively. In its general form, the gate realizes a multi-qubit entangling operation in a single step, whose decomposition into singleand two-qubit gates would require a number of gates that is exponential in the number of qubits. Its demonstrated direct implementation as perceptron in quantum hardware may therefore lead to more powerful quantum neural networks when combined with suitable additional standard gates.

Extensible circuit-QED architecture via amplitude- and frequency-variable microwaves

“…However, direct-digital synthesis using higher-bandwidth AWG's would be a better solution, allowing us to digitally specify the pulse parameters without the need for analog sideband mixing. Certain aspects of implementing frequency chirps in circuit QED have been discussed in the context of fundamental studies of dynamic phaselocking [48][49][50][51], quantum simulation [52] and the realization of a quantum perception [53].…”

A model of interacting quantum neurons with a dynamic synapse

“…It is natural then to model the synaptic depression as an effect between layers. In this direction, there are some network models [20][21][22] and the extension of them with more complicated quantum neuron-neuron interaction would be a very interesting field of research. Besides, recurrent networks, as Hopfield's, are also very interesting and have been recently extended to the quantum regime [23,24].…”

“…In the field of QNN there have been theoretical proposals of models of single quantum neurons [17][18][19], as well as networks such as the perceptron [20][21][22] and Hopfield's [23,24]. The main interest here has been to see if such quantum versions of neural networks are able to improve the properties of classification and pattern recognition compared with classical ones, but there is also a biological motivation [25].…”