Researchers at Paderborn University have developed the Paderborn Quantum Sampler (PaQS), a new platform capable of switching between eight distinct quantum input states within a single experimental run. This architecture allows for direct benchmarking of different sampling regimes, addressing a critical need as the quantum computing community expands beyond initial sampling tasks toward real-world applications. The platform uses a 12-mode interferometer to compare Gaussian and non-Gaussian states, essential resources for outperforming classical systems. This work introduces a hybrid approach to quantify performance costs associated with reducing non-Gaussian resources.
PaQS Platform Enables Switching Between 8 Quantum Input States
This allows for direct comparison of Gaussian and non-Gaussian states within the same 12-mode interferometer, mitigating the influence of long-term drifts that can affect data accuracy. Researchers verified that data generated by PaQS cannot be replicated using classical computational resources, a step toward demonstrating quantum advantage. This verification relies on a semi-device-independent framework based on matrix of moments, which avoids assumptions about the interferometer or input states themselves.
The platform’s design emphasizes integration, with both the squeezed-light source and programmable interferometer realized on integrated photonic platforms, streamlining the process of switching between input states. Discarding the heralding component of data obtained in SBS effectively prepares thermal input states for the interferometer, enabling a comprehensive comparison of different quantum resources. The PaQS platform’s benchmarking framework focuses on certifying the presence of quantumness in generated data, a necessary condition for achieving quantum computational complexity.
Unlike previous approaches that attempt to directly certify computational complexity, this method assesses the fundamental quantum characteristics of the sampled data. The ability to drive the system with single-mode squeezed-vacuum (SMSV) states preserves the computational complexity of the original boson sampling proposal without requiring non-Gaussian inputs. This finding enabled a sequence of increasingly large experiments aimed at demonstrating robust quantum computational advantage, while also opening avenues for exploring real-world applications of sampling architectures.
By implementing squeezed boson sampling (SBS), the second mode of each TMSV input state is directed to photon number resolving (PNR) detectors. The PaQS platform’s modular design and rapid switching capabilities provide a versatile tool for investigating the trade-offs between different quantum resources and optimizing sampling strategies for specific computational tasks.
Boson Sampling Origins & Shift to Gaussian States
This capability addresses a growing need to quantify the performance trade-offs as the field moves from foundational demonstrations toward practical applications. The platform, named the Paderborn Quantum Sampler (PaQS), uses a 12-mode interferometer to switch between single-mode squeezed-vacuum (SMSV) states, heralded Fock states, and thermal states, enabling implementation of Gaussian boson sampling (GBS), squeezed boson sampling (SBS), and thermal boson sampling (TBS). The shift from ideal single-photon inputs to Gaussian states, while easing experimental demands, inherently reduces the non-Gaussianity, a resource for achieving quantum computational advantage, prompting a need to understand the associated performance costs.
The PaQS platform’s design allows for benchmarking these different sampling regimes under otherwise identical conditions, with only the input states varying between configurations. This is particularly important because the type of nonclassicality best suited for a specific task isn’t necessarily correlated with the absolute amount of nonclassicality present in the input states.
The conceptual implementation of various sampling configurations within the PaQS platform involves input fields entering the interferometer from the bottom and being detected with photon-number-resolving (PNR) detectors upon exiting at the top. The ability to systematically compare these states is a conceptual shift in evaluating boson sampling architectures, moving beyond simply maximizing nonclassicality to understanding how different resources contribute to performance in specific computational tasks. The work highlights that the specific behaviors arising from distinct input-state resources require careful consideration as the field matures and explores real-world applications.
PaQS Architecture: 12-Mode Interferometer with Hybrid Inputs
This integrated interferometer offers complete programmability of the implemented unitary transformation, allowing researchers to rapidly reconfigure the system for different sampling regimes without physical adjustments. The PaQS platform’s ability to perform sampling experiments with eight distinct Gaussian or non-Gaussian input states within a single run represents an advancement in quantum sampling technology.
Matrix of Moments Verifies Nonclassicality in PaQS
The Paderborn Quantum Sampler (PaQS) platform verifies nonclassicality through a matrix of moments approach, enabling certification independent of detailed knowledge of the interferometer or input states. This capability allows researchers to confirm the data’s quantum origin, a necessary condition for demonstrating quantum computational advantage, without relying on assumptions about the experimental setup itself. The PaQS design facilitates a semi-device-independent framework, bolstering the robustness of the findings against certain imperfections inherent in real-world experiments.
This direct comparison is achieved while maintaining otherwise identical conditions, ensuring that any observed differences stem solely from the input states employed. Analysis of data generated by PaQS reveals that nonclassical photon-number correlations exist between output modes, confirming quantumness, a prerequisite for genuine quantum computation and efficient simulation resistance.
The researchers use Glauber’s coherence theory, focusing on photon-number covariances between output modes to identify these quantum properties. The study demonstrates advantages for squeezed boson sampling (SBS) in the experimental implementation. A photon-number resolution cutoff of three proves sufficient to verify nonclassicality using the platform’s metric, and, unlike Gaussian boson sampling (GBS), the nonclassicality increases with brighter input states. The team acknowledges that the precise mechanism behind the observed advantages of SBS remains unresolved, but the results emphasise the importance of careful consideration when comparing different sampling regimes.
The ability to rapidly switch between GBS and SBS within PaQS enables a fair benchmarking process, providing valuable insights into the trade-offs between these approaches. The platform’s design and verification methods establish a robust foundation for future investigations into the limits of quantum sampling and the pursuit of quantum computational advantage.
Non-Gaussian Inputs Enhance Nonclassicality Verification
Analysis of datasets produced by the platform demonstrates that nonclassicality, as measured by minimum eigenvalues, steadily increases with brighter inputs in scattershot boson sampling (SBS), while Gaussian boson sampling (GBS) exhibits strong nonclassical signatures only at low brightness levels. This finding challenges the expectation that nonclassicality of squeezed states should increase with greater squeezing, as GBS fails to maintain these signatures as input brightness increases.
This capability is important for quantifying the performance cost associated with reducing non-Gaussian resources, a critical consideration as the field moves beyond initial sampling tasks toward practical applications. This effect is detailed further in an appendix to the study.
The multiplexed detection module attached to one of the platform’s PBS outputs spatially and temporally separates incoming pulses to measure photon number, enabling photon-number heralding for higher-order SBS and verification of the lack of correlations between the two PBS outputs in the GBS configuration. The study emphasises that verification of nonclassicality is a necessary, though not sufficient, condition for quantum computational complexity, highlighting the importance of quantifying these resources for future advancements.
Benchmarking PaQS for Real-World Application Potential
Thermal states are realized by discarding the second mode of each TMSV state, effectively tracing it out. The PaQS system is divided into constituent subsystems for clarity, as detailed in the experimental setup section, and the team characterizes these subsystems with methods described in Appendix B. The benchmarking framework applied to experimental data directly probes for nonclassical photon-number correlations between output modes, identifying quantumness.
This approach rests on the understanding that quantumness is a fundamental requirement for genuine quantum computation and for generating results that defy efficient classical simulation. The nonclassicality criterion used is independent of the implemented unitary transformation, enhancing robustness against certain experimental imperfections.
👉 More information
🗞 Benchmarking Gaussian and Non-Gaussian Input States with a Hybrid Sampling Platform
✍️ Michael Stefszky et al.
🧠 DOI: http://link.aps.org/doi/10.1103/kqyv-lvvy




See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
