[Paper Review] A general-purpose single-photon-based quantum computing platform

Nicolas Maring, Andreas Fyrillas+25 more|arXiv (Cornell University)|Jun 1, 2023

Neural Networks and Reservoir Computing8 citations

TL;DR

The Ascella platform demonstrates a user-ready, general-purpose quantum computer based on six single photons on a reconfigurable 12-mode photonic chip, achieving gate fidelities up to 99.6% for 1-qubit gates, 93.8% for 2-qubit gates, and 86% for 3-qubit gates, plus VQE, photon-native classification, 6-photon Boson Sampling, and heralded 3-photon GHZ generation.

ABSTRACT

Quantum computing aims at exploiting quantum phenomena to efficiently perform computations that are unfeasible even for the most powerful classical supercomputers. Among the promising technological approaches, photonic quantum computing offers the advantages of low decoherence, information processing with modest cryogenic requirements, and native integration with classical and quantum networks. To date, quantum computing demonstrations with light have implemented specific tasks with specialized hardware, notably Gaussian Boson Sampling which permitted quantum computational advantage to be reached. Here we report a first user-ready general-purpose quantum computing prototype based on single photons. The device comprises a high-efficiency quantum-dot single-photon source feeding a universal linear optical network on a reconfigurable chip for which hardware errors are compensated by a machine-learned transpilation process. Our full software stack allows remote control of the device to perform computations via logic gates or direct photonic operations. For gate-based computation we benchmark one-, two- and three-qubit gates with state-of-the art fidelities of $99.6\pm0.1 \%$, $93.8\pm0.6 \%$ and $86\pm1.2 \%$ respectively. We also implement a variational quantum eigensolver, which we use to calculate the energy levels of the hydrogen molecule with high accuracy. For photon native computation, we implement a classifier algorithm using a $3$-photon-based quantum neural network and report a first $6$-photon Boson Sampling demonstration on a universal reconfigurable integrated circuit. Finally, we report on a first heralded 3-photon entanglement generation, a key milestone toward measurement-based quantum computing.

Motivation & Objective

• Demonstrate a general-purpose, user-ready single-photon quantum computing prototype on a reconfigurable chip.
• Show high-fidelity gate-based and photon-native computations on up to six photons.
• Benchmark gate fidelities against existing platforms and demonstrate intermediate-scale quantum tasks (VQE, Boson Sampling, quantum neural network).
• Provide a software stack and cloud access for remote task submission and result retrieval.

Proposed method

• Utilize an on-demand quantum-dot single-photon source delivering six photons at synchronised input modes via an active demultiplexer.
• Interfere photons on a 12-mode Si3N4 universal interferometer with 126 thermo-optic phase shifters and 132 directional couplers.
• Implement a machine-learned transpilation/compilation to compensate hardware errors and map user tasks to the chip controls.
• Characterize and optimize chip control by fitting a thermo-optic model ϕ = A V ⊙2 + b and learning A and b to maximize unitary fidelity.
• Adopt a Perceval-based cloud framework for remote job submission (PC, GB, or U) and collect output coincidences for analysis.
• Perform gate benchmarking via a symmetry-based, finite-sampling fidelity method to obtain Favg(U) for 1-, 2-, and 3-qubit gates.
• Demonstrate a variational quantum eigensolver (VQE) for H2, with classical-quantum feedback optimizing ansatz parameters.
• Explore photon-native computation with a quantum neural network on 5 modes using 3 photons and partial photon-number resolution.
• Execute a 6-photon Boson Sampling experiment on a reconfigurable chip and compare to ideal distributions.
• Show heralded 3-photon GHZ state generation on chip as a milestone toward measurement-based QC.

Experimental results

Research questions

• RQ1 What fidelities can be achieved for 1-, 2-, and 3-qubit gates implemented on Ascella, and how do these compare to other platforms?
• RQ2 Can a six-photon, on-chip photonic processor support gate-based and photon-native computations with high stability and remote operability?
• RQ3 Is it possible to perform non-classical photonic tasks (VQE, quantum neural networks, Boson Sampling, heralded GHZ) on a single integrated platform with cloud access?
• RQ4 How effective is machine-learning-assisted transpilation in compensating fabrication and thermal crosstalk errors to reach high unitary fidelities?

Key findings

• Gate fidelities achieved: 1-qubit T-gate 99.6 ± 0.1%, 2-qubit CNOT 93.8 ± 0.6%, 3-qubit Toffoli 86 ± 1.2%.
• Demonstrated a variational quantum eigensolver for hydrogen (H2) achieving ground-state energy within ±0.01 Hartree after 50–100 iterations.
• On-chip six-photon Boson Sampling with fidelities around 0.97 ± 0.03 and total variation distance ≈ 0.16 ± 0.02.
• Photon-native quantum classifier trained on IRIS data achieved 0.92 training accuracy and 0.95 test accuracy.
• Heralded generation of 3-photon GHZ states on-chip with fidelity 0.82 ± 0.04, marking a key milestone for measurement-based QC.
• Achieved 4 Hz rate for 6-photon processing and long-term stability over weeks, with high photon indistinguishability (~94%).