Superpolynomial Quantum Enhancement in Polaritonic Neuromorphic Computing

TL;DR

This paper predicts a superpolynomial quantum enhancement in image recognition using exciton-polaritons, leveraging their quantum properties within experimentally feasible conditions, advancing neuromorphic computing capabilities.

Contribution

It introduces a theoretical prediction of quantum advantage in polaritonic neuromorphic computing, extending prior classical experiments to the quantum regime.

Findings

Quantum enhancement surpasses classical limits in image recognition

Predicted superpolynomial speedup achievable with current technology

Quantum states of exciton-polaritons enable improved neuromorphic performance

Abstract

Recent proof-of-principle experiments have demonstrated the implementation of neuromorphic computing using exciton-polaritons, making use of coherent classical states [D. Ballarini et al., Nano Lett. 20, 3506 (2020)]. At the same time, it is expected that nonlinear exciton-polaritons can reach a quantum regime forming non-classical states. Here we consider theoretically the quantum nature of exciton polaritons and predict a superpolynomial quantum enhancement in image recognition tasks. This is achieved within experimentally accessible parameters.