Quantum computational advantage with a programmable photonic processor

Abstract

A quantum computer attains computational advantage when outperforming the best classical computers running the best-known algorithms on well-defined tasks. No photonic machine offering programmability over all its quantum gates has demonstrated quantum computational advantage: previous machines1,2 were largely restricted to static gate sequences. Earlier photonic demonstrations were also vulnerable to spoofing3, in which classical heuristics produce samples, without direct simulation, lying closer to the ideal distribution than do samples from the quantum hardware. Here we report quantum computational advantage using Borealis, a photonic processor offering dynamic programmability on all gates implemented. We carry out Gaussian boson sampling4 (GBS) on 216 squeezed modes entangled with three-dimensional connectivity5, using a time-multiplexed and photon-number-resolving architecture. On average, it would take more than 9,000 years for the best available algorithms and supercomputers to produce, using exact methods, a single sample from the programmed distribution, whereas Borealis requires only 36 μs. This runtime advantage is over 50 million times as extreme as that reported from earlier photonic machines. Ours constitutes a very large GBS experiment, registering events with up to 219 photons and a mean photon number of 125. This work is a critical milestone on the path to a practical quantum computer, validating key technological features of photonics as a platform for this goal.

Computational advantage of quantum random sampling

“…Similarly, we can also think of the reverse, where a photon-number state is prepared in the input and Gaussian measurements are performed (Chakhmakhchyan and Cerf, 2017). Those Gaussian boson sampling protocols are appealing in comparison to the original proposal as Gaussian states and measurements are experimentally much easier to implement than photon-number states and measurements, and indeed it is those protocols for which large-scale experiments have been performed recently (Madsen et al, 2022;Wang et al, 2019;Zhong et al, 2021Zhong et al, , 2020.…”

“…At the same time, there is significant evidence that current-day supercomputers have a very hard time simulating this task even for small systems of size roughly 50 to 100 (Bulmer et al, 2021;Huang et al, 2020b;Markov et al, 2018;Neville et al, 2017;. Very recently, quantum random sampling in a classically intractable regime has been achieved experimentally on a universal quantum processor comprising 53 qubits (Arute et al, 2019), and up to 60 qubits (Wu et al, 2021;Zhu et al, 2022), as well as using photonic systems (Madsen et al, 2022;Zhong et al, 2021Zhong et al, , 2020.…”

Gaussian Boson Sampling Compilation

“…It is in these operands where the nature of the computation resides, and what links the combinatorics problem to the calculation of the so called permanent of the operand, as it will be shown in Section 3. The work by Madsen et al [5] more recently claimed quantum advantage with a programmable photonic processor, a significant advancement towards utilization of GBS.…”

“…This platform is now suited to solve small size problems (up to 4x4 size matrices) on their X8 hardware platform, through the application of Gaussian operands. Despite the claimed quantum advantage [5], this is an emerging effort, still not ready for general comparison against classical implementations. To reach the point in which continuous variable implementations are ready to challenge the performance of classical implementations, two key advances need to be in place: (i) the handling of larger size problems, and (ii) higher flexibility in the kinds of operands that can be applied to the initial quantum state of the bosonic system.…”

Measurement-Based Quantum Walks on High-Dimensional Graphs

“…Additionally, our platform is compatible with the inclusion of single-photon states [74][75][76], thus rendering it possible for the implementation of multimode interference of multiphoton states in large networks that lies at the heart of next-generation quantum information applications [77]. Finally, the possibility to include an appropriate parametric down-conversion source at the input along with a photon-number-resolved detection of the up-converted output light of the QPG facilitates the scalable realization of fully integrated Gaussian Boson Sampling [78], which is the first photonic system to claim a quantum advantage and a nearterm platform for photonic quantum computation [79][80][81].…”