Large-scale quantum reservoir computing using a Gaussian Boson Sampler
Valeria Cimini, Mandar M. Sohoni, Federico Presutti, Benjamin K. Malia, Shi-Yuan Ma, Ryotatsu Yanagimoto, Tianyu Wang, Tatsuhiro Onodera, Logan G. Wright, Peter L. McMahon
Abstract A Gaussian boson sampler (GBS) is a special-purpose quantum computer that can be practically realized at a large scale in optics. Here we report on experiments in which we used a frequency-multiplexed GBS with > 400 modes as a quantum reservoir. We evaluated the accuracy of our GBS-based reservoir computer on a variety of benchmark tasks. We found that when the system was given access to the correlations between measured modes of the GBS, the achieved accuracies were the same or higher than when it was only given access to the mean photon number in each mode—and in several cases the advantage in accuracy from using the correlations was greater than 20 percentage points. This provides experimental evidence in support of theoretical predictions that access to correlations enhances the power of quantum reservoir computers. We also tested our reservoir computer when operating the reservoir with various sources of classical rather than quantum light and found that using squeezed light consistently resulted in the highest accuracies. Our work experimentally establishes that a GBS can be an effective quantum reservoir and provides a practical platform for experimentally exploring the role of quantumness and correlations in quantum machine learning at very large system sizes.
Abstract A Gaussian boson sampler (GBS) is a special-purpose quantum computer that can be practically realized at a large scale in optics. Here we report on experiments in which we used a frequency-multiplexed GBS with > 400 modes as a quantum reservoir. We evaluated the accuracy of our GBS-based reservoir computer on a variety of benchmark tasks. We found that when the system was given access to the correlations between measured modes of the GBS, the achieved accuracies were the same or higher than when it was only given access to the mean photon number in each mode—and in several cases the advantage in accuracy from using the correlations was greater than 20 percentage points. This provides experimental evidence in support of theoretical predictions that access to correlations enhances the power of quantum reservoir computers. We also tested our reservoir computer when operating the reservoir with various sources of classical rather than quantum light and found that using squeezed light consistently resulted in the highest accuracies. Our work experimentally establishes that a GBS can be an effective quantum reservoir and provides a practical platform for experimentally exploring the role of quantumness and correlations in quantum machine learning at very large system sizes.
Reservoir models offer a hardware-efficient learning paradigm for noisy intermediate-scale quantum devices by exploiting untrained quantum dynamics as a fixed feature map and restricting optimization to a simple classical readout layer. We propose a quantum extreme learning machine implemented using gaussian boson sampling and an encoding strategy that achieves high classification accuracy while reducing optical resource requirements. Classical inputs are jointly encoded in the squeezing parameters and in the interferometer unitary, enabling sampling-based, highly nonlinear feature maps while leveraging large-scale GBS output statistics, which are conjectured to be classically intractable. We systematically compare multiple families of quantum features accessible in the same setup and find that photon-number sampling probabilities provide the best performance, consistent with their higher effective feature dimensionality. Finally, we benchmark against classical nonlinear baselines and analyse robustness under noisy scenarios, showing competitive performance with fewer trainable parameters and indicating practical promise for near-term photonic implementations.
Gaussian boson sampling (GBS) is a quantum sampling task in which one has to draw samples from the photon-number distribution of a large-dimensional nonclassical squeezed state of light.In an effort to make this task intractable for a classical computer, experiments building GBS machines have mainly focused on increasing the dimensionality and squeezing strength of the nonclassical light.However, no experiment has yet demonstrated the ability to displace the squeezed state in phase space, which is generally required for practical applications of GBS.In this work, we build a GBS machine that achieves the displacement by injecting a laser beam alongside a two-mode squeezed vacuum state into a 15-mode interferometer.We focus on two new capabilities.Firstly, we use the displacement to reconstruct the multimode Gaussian state at the output of the interferometer.Our reconstruction technique is in situ and requires only three measurement settings regardless of the state dimension.Secondly, we study how the addition of classical laser light in our GBS machine affects the complexity of sampling its output photon statistics.We introduce and validate approximate semiclassical models that reduce the computational cost when a significant fraction of the detected light is classical.
Photonics is a promising platform for demonstrating a quantum computational advantage (QCA) by outperforming the most powerful classical supercomputers on a well-defined computational task. Despite this promise, existing proposals and demonstrations face challenges. Experimentally, current implementations of Gaussian boson sampling (GBS) lack programmability or have prohibitive loss rates. Theoretically, there is a comparative lack of rigorous evidence for the classical hardness of GBS. In this work, we make progress in improving both the theoretical evidence and experimental prospects. We provide evidence for the hardness of GBS, comparable to the strongest theoretical proposals for QCA. We also propose a QCA architecture we call high-dimensional GBS, which is programmable and can be implemented with low loss using few optical components. We show that particular algorithms for simulating GBS are outperformed by high-dimensional GBS experiments at modest system sizes. This work thus opens the path to demonstrating QCA with programmable photonic processors.
Quantum computers solve intractable problems which classically require an exponentially long time to compute. With the development of large-scale experiments that claim quantum advantage, a vital issue has now emerged. What are the errors, and how do they affect the complexity of the problem solved? Large-scale Gaussian boson sampling (GBS) experiments give an example in which random numbers are generated. Despite classical hardness, these have computable benchmarks for checking data validity. While there are other quantum computing architectures, Gaussian boson sampling is uniquely testable at all scales. Several large, pioneering quantum computing (QC) experiments have been carried out to investigate quantum advantage. Here, we introduce a highly scalable but classical algorithm that can solve GBS approximately. Our numerical simulation of the output count data is closer to the exact solution than current experiments up to 1152 modes. This algorithm outperforms all previous classical, approximate algorithms and scales efficiently to larger experiments. Our results show that effects beyond losses can cause the errors that allow classical simulability. This work will lead to more precise algorithms and is a step towards understanding how QC quantum advantage is affected by the underlying physics.
Quantum machine learning stands poised as a forefront application for near-term quantum devices, addressing scalability challenges posed by classical computers in handling large datasets. Gaussian boson sampling (GBS), an intricate quantum algorithm deemed computationally infeasible for classical counterparts, represents a substantial leap forward in computational tasks. However, to date, the benefits of GBS-assisted quantum unsupervised machine learning are not experimentally demonstrated. Here, we present the first experimental implementation of quantum unsupervised machine learning using the GBS protocol with a universal programmable integrated photonic chip. The experimental system contains 16 squeezing sources, a universal programmable unitary matrix network of 16 modes, and a multi-channel single-photon detector, producing substantial output data crucial for 2 typical types of unsupervised tasks: feature extraction and generative network. Compared to classical approaches, the study demonstrates quantum-enhanced capability in extracting complex features from high-dimensional spaces and improved performance in generating arbitrary curve points and reconstructing handwritten digit images. This work not only underscores the potential of GBS in expressing high-dimensional features but also charts a path toward practical implementations within scalable, dimension-enhanced quantum unsupervised machine learning frameworks. The quantum unsupervised machine learning paradigm, offering theoretical acceleration and reduced training parameters for high-dimensional datasets, shows significant promise for advancing real-world applications of quantum technologies.
Abstract An important challenge with the current generation of noisy, large-scale quantum computers is the question of validation. Does the hardware generate correct answers? If not, what are the errors? This issue is often combined with questions of computational advantage, but it is a fundamentally distinct issue. In current experiments, complete validation of the output statistics is generally not possible because it is exponentially hard to do so. Here, we apply phase-space simulation methods to partially verify recent experiments on Gaussian boson sampling (GBS) implementing photon-number resolving detectors. The positive-P phase-space distribution is employed, as it uses probabilistic sampling to reduce complexity. It is ∼ 10 18 times faster than direct classical simulation for experiments on 288 modes where quantum computational advantage is claimed. When combined with binning and marginalization to improve statistics, multiple validation tests are efficiently computable, of which some tests can be carried out on experimental data. We show that the data as a whole has discrepancies with theoretical predictions for perfect squeezing. However, a modification of the GBS parameters greatly improves agreement for some tests. We suggest that such validation tests could form the basis of feedback methods to improve GBS experiments.