State transfer with separable optical beams and variational quantum algorithms with classical light

TL;DR

This paper demonstrates how classical optical beams can simulate quantum information transfer and machine learning algorithms, enabling quantum-like processing with classical light through novel beam classes and protocols.

Contribution

It introduces equivalent optical beams for information transfer and shows how quantum machine learning can be implemented with classical light, bridging classical optics and quantum algorithms.

Findings

Equivalent optical beams can transfer information between degrees of freedom.

Quantum classifiers can be realized using classical light's orbital angular momentum.

Protocols for experimental implementation are proposed.

Abstract

Classical electromagnetic fields and quantum mechanics -- both obey the principle of superposition alike. This opens up many avenues for simulation of a large variety of phenomena and algorithms, which have hitherto been considered quantum mechanical. In this paper, we propose two such applications. In the first, we introduce a new class of beams, called equivalent optical beams, in parallel with equivalent states introduced in [Bharath & Ravishankar, https://doi.org/10.1103/PhysRevA.89.062110]. These beams have the same information content for all practical purposes. Employing them, we show how to transfer information from one degree of freedom of classical light to another, without any need for classically entangled beams. Next, we show that quantum machine learning can be performed with OAM beams through the implementation of a quantum classifier circuit. We provide explicit protocols and explore the possibility of their experimental realization.