Environment-Assisted Decoherence Suppression of Optical Non-Gaussian States
Akihiro Machinaga, Naoki Aritomi, Ryoga Sakurada, Daichi Okuno, Keitaro Anai, Takahiro Kashiwazaki, Takeshi Umeki, Shigehito Miki
Abstract
Optical loss is a common bottleneck in photonic quantum information processing, undermining the quantum advantage over classical approaches. Although several countermeasures, such as quantum distillation and error correction, have been proposed, they typically require experimentally demanding non-Gaussian operations. Here, we demonstrate a Gaussian-only scheme that suppresses loss-induced decoherence for general, unknown optical quantum states. By injecting a squeezed vacuum state into an environment of the loss channel and performing feedforward based on environmental monitoring, the scheme effectively suppresses loss-induced noise. Our programmable loop-based optical circuit allows us to implement the scheme for several types of loss-sensitive non-Gaussian states under various loss conditions for up to five steps, and directly compare the results with the unsuppressed case. Our results show that the scheme consistently mitigates state degradation, preserving higher fidelity and Wigner negativity than without suppression. This approach can be applied to mitigating a broad class of errors in optical systems and extending quantum memory lifetimes. Moreover, it is compatible with other loss-suppression techniques and extendable to physical platforms beyond optics, offering a promising route toward reducing the overhead required for fault-tolerant quantum information processing.
AI Impact Assessments
(3 models)Scientific Impact Assessment
Core Contribution
This paper demonstrates an environment-assisted decoherence suppression (EADS) scheme that mitigates photon-loss-induced decoherence for general, unknown optical quantum states using only Gaussian operations. The key innovation is replacing the vacuum state entering the loss channel's environment with a squeezed vacuum state, then monitoring the leaked mode via homodyne detection and applying feedforward displacement to cancel added noise. The critical distinction from prior work is that this scheme (a) requires no non-Gaussian elements (unlike distillation and error correction), (b) works for unknown states (unlike preprocessing-based approaches that require knowledge of the input), and (c) operates on the environment rather than the state itself.
The scheme is based on prior theoretical proposals by Marek & Filip (2004) and Zhang, Zou & Jiang (2018), but this work provides the first experimental demonstration with non-Gaussian states, multi-step operation, and systematic benchmarking.
Methodological Rigor
The experimental implementation is well-designed. The authors use a programmable loop-based optical circuit operating at 1545 nm, which elegantly avoids the scaling problem of cascading beam splitters for multi-step implementations. This time-domain multiplexing approach is both practical and clever.
The experimental verification is systematic: three input state types (single-photon, x-squeezed single-photon, p-squeezed single-photon) × two loss conditions (5% and 10% per step) × up to five steps, all compared with unsuppressed baselines. This provides 6 distinct scenarios tested over multiple steps, giving confidence in the generality of the results.
Two concerns reduce rigor slightly: (1) The feedforward is applied numerically in post-processing rather than in real-time, which is a significant caveat for practical deployment. (2) The theoretical model uses fitting parameters (ηNG, ηloop) derived from unsuppressed data and then applied to predict suppressed performance — while reasonable, this introduces some circularity. (3) The internal loop loss (~6%) remains unsuppressed, meaning the demonstrated improvement is partial and the scheme is not self-contained.
The data show some fluctuations relative to theoretical predictions (acknowledged by the authors as arising from phase-locking offsets and gain mismatches), and at small step numbers (N=0,1), the suppressed case actually shows slightly lower fidelity due to imperfect optical switching — an honest and important disclosure.
The ancillary squeezing level of 4.5 dB measured (9.7 dB estimated pure) is substantial but not extreme, and the gap between experimental performance and the infinite-squeezing limit shows clear room for improvement with better squeezing resources.
Potential Impact
Immediate applications: The scheme is directly relevant to extending quantum memory lifetimes in loop-based architectures — a critical need for time-multiplexed quantum computing and for synchronizing probabilistic resource state generation. The compatibility with GKP error correction codes is particularly compelling, as reducing the effective physical loss rate could lower fault-tolerance thresholds.
Broader applicability: The authors convincingly argue the scheme extends beyond pure photon loss to mode mismatch, mode crosstalk, and imperfect quantum transduction — any scenario where the environmental mode is physically accessible. The extension to superconducting microwave circuits and trapped-ion vibrational modes broadens the potential impact significantly.
Practical limitations: The requirement for physical access to the environment is a fundamental constraint. For fiber transmission losses or absorption, the scheme is inapplicable. This limits the communication applications somewhat, though it remains highly relevant for on-chip and local processing losses.
Timeliness & Relevance
This work is highly timely. The photonic quantum computing community is actively grappling with the loss problem as architectures scale. Time-multiplexed approaches (e.g., Xanadu, RIKEN/UTokyo cluster-state generation) face exactly the quantum memory lifetime challenges this scheme addresses. The Gaussian-only requirement is crucial — non-Gaussian operations remain the most experimentally challenging aspect of CV quantum information, so avoiding them dramatically lowers the implementation barrier.
The work also arrives at a moment when GKP-based fault-tolerant architectures are gaining traction in both photonic and superconducting platforms, making the demonstrated compatibility with such codes particularly relevant.
Strengths
1. Gaussian-only implementation: No non-Gaussian resources needed for the suppression itself — a major practical advantage over distillation/error correction approaches.
2. State-agnostic operation: Works for unknown states, unlike preprocessing methods requiring state knowledge.
3. Systematic experimental validation: Six distinct scenarios with multi-step operation provide strong evidence of generality.
4. Programmable platform: The loop-based circuit enables flexible reconfiguration, and the approach directly demonstrates quantum memory lifetime extension.
5. Clear theoretical framework: The Supplementary Information provides a complete Gaussian channel model with analytical expressions for the output Wigner function.
6. Open data and code: Enhances reproducibility.
Limitations
1. Post-processing feedforward: Real-time feedforward was not demonstrated, which is necessary for practical deployment and adds latency/noise considerations.
2. Unsuppressed internal losses: The ~6% loop loss per round trip is not addressed by the scheme, causing continued degradation even with perfect ancilla squeezing.
3. Asymmetric noise suppression: The scheme suppresses noise in only one quadrature direction at a time, introducing a deterministic squeezing byproduct that must be compensated.
4. Environmental access requirement: Fundamentally limits applicability to scenarios where the loss environment is physically accessible.
5. Modest absolute improvement: While the relative improvement is clear, the absolute fidelity and negativity values remain limited by finite squeezing and loop losses, and the advantage narrows for larger step numbers.
6. No demonstration with higher photon-number states: Testing was limited to single-photon-based states; more complex non-Gaussian states (cat states, multi-photon Fock states) would strengthen the generality claim.
Overall Assessment
This is a solid experimental demonstration of an important theoretical concept, executed with good systematic rigor on a versatile platform. While the improvements are incremental in absolute terms and the scheme has fundamental limitations (environmental access, asymmetric noise), its Gaussian-only nature and state-agnostic operation fill a genuine gap in the toolbox for loss mitigation. The work is more impactful as a proof-of-principle and platform demonstration than as an immediately deployable technology.
Generated Apr 9, 2026
Comparison History (73)
Paper 2 offers a platform-agnostic, decoder-independent method to significantly improve quantum error correction thresholds without hardware modifications. Its broad applicability across all fault-tolerant quantum computing architectures and substantial reduction in logical error rates give it a wider and higher potential impact compared to Paper 1, which is focused specifically on continuous-variable optical systems.
Paper 1 presents an experimental breakthrough addressing optical loss, a major physical bottleneck in quantum information processing. Its Gaussian-only scheme provides a practical, demonstrated solution for suppressing decoherence that extends to broad classes of errors and other physical platforms. Paper 2 offers a valuable theoretical tool for simulating surface codes, but Paper 1's experimental validation and potential to directly reduce overhead for fault-tolerant quantum computing give it broader and more immediate scientific impact.
Paper 2 addresses a fundamental bottleneck in quantum information processing—optical loss and decoherence—by providing an experimentally validated, broadly applicable scheme. Its potential to extend quantum memory lifetimes and reduce overhead for fault-tolerant quantum computing offers broader real-world applications and cross-disciplinary impact compared to Paper 1, which primarily focuses on a methodological improvement for numerical simulations in computational physics.
Paper 1 likely has higher impact: it experimentally demonstrates a Gaussian-only, programmable scheme to suppress loss-induced decoherence for unknown non-Gaussian optical states, preserving fidelity and Wigner negativity. This is a timely, practically enabling advance for photonic quantum computing and quantum memories, potentially reducing fault-tolerance overhead and generalizing beyond optics. Paper 2 offers valuable, rigorous analytical/algorithmic approximations for surface-code logical error rates, improving design-space exploration, but its assumptions (i.i.d. noise) and primarily computational contribution may limit breadth compared with Paper 1’s cross-platform, experimentally validated error-mitigation technique.
Paper 1 likely has higher impact: it experimentally demonstrates a Gaussian-only, environment-assisted feedforward scheme to suppress optical-loss decoherence for unknown non-Gaussian states, addressing a central bottleneck in photonic quantum tech with clear near-term applications (quantum memories, fault-tolerant overhead reduction) and potential platform generality. The work appears methodologically strong (programmable multi-step implementation, direct benchmarks, Wigner-negativity preservation). Paper 2 is a valuable methodological improvement for time-dependent VMC, but its impact is narrower and partially exploratory (TCI path not yet successful), and adoption depends on community uptake and further validation.
Paper 2 addresses optical loss, a critical bottleneck in photonic quantum computing, with an experimentally demonstrated Gaussian-only scheme. Its practical approach to suppressing decoherence without demanding non-Gaussian operations offers broad, immediate applications for extending quantum memory and advancing fault-tolerant architectures. While Paper 1 presents an innovative algorithmic approach to circuit design, Paper 2 has a more direct and profound impact on overcoming fundamental physical limitations in quantum hardware.
Paper 2 demonstrates a more broadly applicable technique—a Gaussian-only scheme for suppressing decoherence in optical quantum systems that works for general unknown states without demanding non-Gaussian operations. This significantly lowers experimental overhead for loss mitigation, a universal bottleneck in photonic quantum computing. Its compatibility with other techniques and extensibility beyond optics gives it broader cross-platform impact. Paper 1, while clever and practical for superconducting qubits with TLS defects, addresses a more specific problem with a narrower scope of applicability.
Paper 2 addresses a critical scalability bottleneck in fault-tolerant quantum computing—real-time decoding resource management for qLDPC codes. Its automated predecoder framework achieving 3,963x decoder utilization reduction and concrete hardware implementations (FPGA/ASIC) supporting hundreds of thousands of logical qubits represent highly practical advances. The work bridges an important gap as the field shifts from surface codes to qLDPC codes. Paper 1 demonstrates useful decoherence suppression but addresses a narrower problem with incremental improvements. Paper 2's broader architectural impact on the quantum computing stack and timeliness with the qLDPC transition give it higher potential impact.
Paper 1 addresses a fundamental bottleneck in quantum computing—decoherence and optical loss—with an experimentally demonstrated, Gaussian-only mitigation scheme. Its ability to extend quantum memory lifetimes and reduce overhead for fault-tolerant quantum information processing gives it broad, immediate real-world applicability. In contrast, Paper 2 provides important theoretical constructions for unitary k-designs on grids, which advances foundational quantum circuit theory but has narrower immediate practical impact compared to solving physical decoherence challenges in scalable quantum hardware.
Paper 1 likely has higher impact: it experimentally demonstrates a Gaussian-only, programmable, multi-step protocol that suppresses loss-induced decoherence for unknown non-Gaussian optical states while preserving Wigner negativity—directly addressing a key bottleneck in photonic quantum information with near-term applicability and potential to reduce fault-tolerance overhead. Paper 2 is a strong and rigorous theoretical framework for cQED noise-model reduction, but its impact depends on adoption and validation against complex hardware; it is more specialized and less immediately transformative across platforms than an experimentally validated loss-suppression method.