Quantum Light Detection with Enhanced Photonic Neural Network

TL;DR

This work tackles the bottleneck of weak optical nonlinearities in bosonic photonic quantum reservoirs for sensing. It introduces a hybrid quantum–classical architecture (EQSS) that couples a bosonic reservoir to a classical FFNN readout, achieving superior state classification, regression, and tomography compared with standard QRC while reducing reliance on strong nonlinearity or large reservoirs. Demonstrated gains include classification accuracy improvements from ~0.78 to >0.96 and order-of-magnitude reductions in regression error, with effective operation at $U/\gamma$ as low as $0.02$ in a five-node reservoir. The results point to a scalable, chip-ready sensing platform that can be integrated with existing photonic quantum systems, highlighting a practical route toward adaptive, learning-based quantum sensors that leverage both quantum nonlinear dynamics and classical processing.

Abstract

Advances in quantum technologies are accelerating the demand for optical quantum state sensors that combine high precision, versatility, and scalability within a unified hardware platform. Quantum reservoir computing offers a powerful route toward this goal by exploiting the nonlinear dynamics of quantum systems to process and interpret quantum information efficiently. Photonic neural networks are particularly well suited for such implementations, owing to their intrinsic sensitivity to photon-encoded quantum information. However, the practical realisation of photonic quantum reservoirs remains constrained by the inherently weak optical nonlinearities of available materials and the technological challenges of fabricating densely coupled quantum networks. To address these limitations, we introduce a hybrid quantum-classical detection protocol that integrates the advantages of quantum reservoirs with the adaptive learning capabilities of analogue neural networks. This synergistic architecture substantially enhances information-extraction accuracy and robustness, enabling low-cost performance improvements of quantum light sensors. Based on the proposed approach, we achieved significant improvements in quantum state classification, tomography, and feature regression, even for reservoirs with a relatively small nonlinearity-to-losses ratio $U/γ\approx 0.02$ in a network of only five nodes. By reducing reliance on material nonlinearity and reservoir size, the proposed approach facilitates the practical deployment of high-fidelity photonic quantum sensors on existing integrated platforms, paving the way toward chip-scale quantum processors and photonic sensing technologies.

Figures (2)

• Figure 1: Enhanced photonic reservoir neural network for multi-task quantum state sensing.a The sequence of probing quantum states and coherent laser field acting on the optical resonator. An input quantum state drives the dynamics of an optical reservoir formed by coupled bosonic modes within a cavity, sequentially. b Time-dependent occupations of the reservoir nodes, encoding information about the input state’s density matrix, are continuously monitored by a detection system (schematic illustration). c The recorded reservoir responses are processed by a classical feed-forward neural network, implemented either in software or hardware, which is trained to extract task-relevant quantum-state features from nonlinear reservoir dynamics. d The hybrid sensor performs multiple quantum sensing tasks, including state classification, parameter regression, and quantum state tomography.
• Figure 2: Enhanced optical reservoir network as a multitask quantum state sensor.a, Results of quantum state classification, single- and multi-parameter feature regression, and quantum state tomography obtained using the standard QRC approach. b, Corresponding results achieved with the EQSS detection strategy. Each panel in (b) directly corresponds to its counterpart in (a), sharing identical reservoir parameters and datasets. c,d, Dependence of classification accuracy and mean-squared error on reservoir nonlinearity $U$ and size $N$ for each benchmark task (columns correspond to specific tasks). Blue circles and orange squares represent results from EQSS and QRC, respectively, while grey circles indicate EQSS performance with a linear reservoir ($U=0$). Each data point was averaged over 10 independent training–testing realisations, with error bars denoting the standard deviation. The last row insets in panels a and b, represent the target state for quantum tomography task.