pith. sign in

arxiv: 2605.02045 · v1 · submitted 2026-05-03 · 💻 cs.IT · cs.AI· math.IT· quant-ph

Optimization of CV-QKD Under Practical Constraints

Svitlana Matsenko , Amirhossein Ghazisaeidi , Marcin Jarzyna , Konrad Banaszek , Darko Zibar This is my paper

Pith reviewed 2026-05-08 18:53 UTC · model grok-4.3

classification 💻 cs.IT cs.AImath.ITquant-ph
keywords CV-QKDreinforcement learningpractical constraintsFIR filtersDAC ADC resolutionsecret key ratequantum key distribution
0
0 comments X

The pith

Reinforcement learning optimizes CV-QKD performance under practical hardware constraints like limited filter taps and finite bit resolution.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This paper shows how reinforcement learning can adjust the settings of a continuous-variable quantum key distribution system to work better when hardware has real-world limits. The constraints considered are the number of taps in FIR filters, the average number of photons, and the precision of DAC and ADC converters. If successful, this would mean quantum key distribution can achieve higher rates and better security in actual deployed systems rather than only in perfect lab conditions. Readers would care because it makes theoretical quantum cryptography more relevant to engineering practice by accounting for imperfections that reduce performance.

Core claim

Using reinforcement learning, we optimize for practical hardware constraints, including limited FIR filter taps at the transmitter and receiver, mean photon number and finite DAC/ADC resolution. Under these realistic conditions, the proposed approach achieves significant performance improvements.

What carries the argument

A reinforcement learning agent that selects CV-QKD system parameters to maximize the secret key rate while respecting the hardware limits on filters, photon number, and converter resolution.

If this is right

  • Higher secret key rates become achievable in CV-QKD links that must use limited FIR filter taps at both ends.
  • Finite DAC and ADC resolution no longer impose as large a penalty on the achievable rate when parameters are RL-tuned.
  • The mean photon number can be chosen to balance rate and security more effectively under the other hardware limits.
  • Overall system performance improves without requiring upgrades to filter complexity or converter bit depth.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same reinforcement learning approach could be extended to optimize other quantum communication protocols that face similar hardware constraints.
  • Integrating the trained agent with real-time channel monitoring might allow ongoing adaptation in deployed quantum networks.
  • Laboratory tests with actual hardware would directly confirm whether the simulated gains translate to physical implementations.

Load-bearing premise

That the reinforcement learning procedure can be trained and evaluated under the stated constraints in a way that produces genuine, reproducible gains rather than artifacts of the simulation or training setup.

What would settle it

A physical CV-QKD experiment that measures the secret key rate achieved with the RL-optimized parameters against the rate from a conventional optimization under identical filter tap, photon number, and resolution limits.

read the original abstract

Using reinforcement learning, we optimize for practical hardware constraints, including limited FIR filter taps at the transmitter and receiver, mean photon number and finite DAC/ADC resolution. Under these realistic conditions, the proposed approach achieves significant performance improvements.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript proposes using reinforcement learning to optimize continuous-variable quantum key distribution (CV-QKD) systems subject to practical hardware constraints, specifically limited FIR filter taps at transmitter and receiver, mean photon number, and finite DAC/ADC resolution. It claims that this yields significant performance improvements relative to unoptimized or conventionally optimized systems under realistic conditions.

Significance. If the performance gains are substantiated by reproducible simulations with appropriate baselines, statistical error bars, and clear reward-function definitions, the work would be significant for bridging theoretical CV-QKD rate calculations with hardware realities, potentially informing practical system design and increasing achievable secure key rates in constrained deployments.

major comments (2)
  1. Abstract: the central claim of 'significant performance improvements' is unsupported by any quantitative metrics, key-rate values, baseline comparisons, or error analysis, which is load-bearing for the paper's contribution and prevents verification of the result.
  2. No methods or results sections are available to inspect the RL formulation (state/action spaces, reward function, training procedure), simulation model for the CV-QKD channel and hardware constraints, or the specific FIR-tap and resolution values used; without these the reproducibility of any reported gains cannot be assessed.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major comment below and have revised the manuscript to strengthen the presentation of our results and improve reproducibility.

read point-by-point responses

  1. Referee: Abstract: the central claim of 'significant performance improvements' is unsupported by any quantitative metrics, key-rate values, baseline comparisons, or error analysis, which is load-bearing for the paper's contribution and prevents verification of the result.

    Authors: We agree that the abstract requires quantitative support to substantiate the claim. We have revised the abstract to incorporate specific key-rate values and relative improvements from our simulations, along with references to the baseline methods and the statistical reliability of the results as detailed in the Results section. revision: yes

  2. Referee: No methods or results sections are available to inspect the RL formulation (state/action spaces, reward function, training procedure), simulation model for the CV-QKD channel and hardware constraints, or the specific FIR-tap and resolution values used; without these the reproducibility of any reported gains cannot be assessed.

    Authors: The manuscript contains dedicated sections describing the RL formulation and simulation results. However, we acknowledge that the level of detail may not have been sufficient for full reproducibility. We have expanded the Methods section with explicit definitions of the state and action spaces, the reward function, training procedure, the CV-QKD channel and hardware model, and a table of the specific parameter values (FIR taps, photon number, DAC/ADC resolution). The Results section has also been augmented with additional simulation details and error analysis. revision: partial

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The provided abstract and description present a standard application of reinforcement learning to optimize CV-QKD parameters under hardware constraints such as FIR filter taps, mean photon number, and DAC/ADC resolution. No equations, self-definitional loops, fitted inputs renamed as predictions, or load-bearing self-citations appear in the text. The central claim of performance improvements is framed as an empirical outcome of the RL procedure rather than a mathematical derivation that reduces to its own inputs by construction. Without any visible derivation chain or ansatz smuggling, the approach remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities are described in the abstract. The approach is presented as standard reinforcement learning applied to an existing problem.

pith-pipeline@v0.9.0 · 5335 in / 985 out tokens · 22289 ms · 2026-05-08T18:53:38.109091+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

11 extracted references · 11 canonical work pages

  1. [1]

    ), (1) namely the mean photon number of the transmit-ted signal 𝑛*, the effective system transmittance 𝜏, and the effective system excess noise 𝑛!

    Optimization of CV-QKD Under Practical Constraints Svitlana Matsenko(1), Amirhossein Ghazisaeidi(2), Marcin Jarzyna(3), Konrad Banaszek(3,4), and Darko Zibar(1) (1) DTU Electro, Technical University of Denmark, DK-2800, Kgs. Lyngby, Denmark, svitma@dtu.dk (2) Nokia Bell Labs, 91300 Massy, France (3) Centre for Quantum Optical Technologies, CeNT, Universit...

    work page 2026

  2. [2]

    Acknowledgements M.J. and K.B. acknowledge support by the Euro-pean Union’s Horizon Europe research and inno-vation programme under the project ‘Quantum Security Networks Partnership’ (QSNP, Grant Agreement No. 101114043) and the ‘Quantum Optical Technologies’ project (FENG.02.01-IP.05-0017/23) carried out within the Interna-tional Research Agendas progra...

    work page 2021

  3. [3]

    Zhang, Y

    Y. Zhang, Y. Bian, Z. Li, S. Yu, and H. Guo, “Continuous-variable quantum key distribution system: Past, present, and future”, Applied Physics Reviews, vol. 11, no. 1, Mar. 2024, ISSN: 1931-9401, DOI: 10.1063/5.0179566

    work page doi:10.1063/5.0179566 2024

  4. [4]

    Automatic mitigation of dynamic atmospheric turbulence using optical phase conjugation for coher- ent free-space optical communications

    H. Wang, Y. Li, et al., “High-rate continuous-variable quantum key distribution over 100 km fiber with compos-able security”, Optica, vol. 12, no. 10, pp. 1657–1667, Oct. 2025, ISSN: 2334-2536, DOI: 10.1364/OPTICA. 566359

    work page doi:10.1364/optica 2025

  5. [5]

    Wavelength division multiplexing of continuous variable quantum key distribution and 18.3 Tbit/s data channels

    T. A. Eriksson et al., “Wavelength division multiplexing of continuous variable quantum key distribution and 18.3 Tbit/s data channels”, Communications Physics, vol. 2, no. 9, Jan. 2019, DOI: 10.1038/s42005-018-0105-5

    work page doi:10.1038/s42005-018-0105-5 2019

  6. [6]

    Unconditional Security Proof of Long-Distance Continuous-Variable Quantum Key Distribution with Discrete Modulation

    A. Leverrier and P. Grangier, “Unconditional security proof of long-distance continuous-variable quantum key distribution with discrete modulation”, Physical Review Letters, vol. 102, no. 18, May 2009, DOI: 10.1103/PhysRevLett.102.180504

    work page doi:10.1103/physrevlett.102.180504 2009

  7. [7]

    430 Tb/S GMI Data Rate Over a Stan- dard G.654 Fiber Using Few-Mode O-Band and Single- Mode ESCL-Band Transmission

    S. Matsenko et al., “Mode mismatch mitigation in Gauss-ian-modulated CV-QKD”, in 2025 European Conference on Optical Communications (ECOC), 2025, pp. 1–4, DOI: 10.1109/ECOC66593.2025.11263288

    work page doi:10.1109/ecoc66593.2025.11263288 2025

  8. [8]

    Tx-Rx mode mismatch effects in gaussian-modulated CV-QKD

    M. Kucharczyk, M. Jachura, M. Jarzyna, K. Banaszek, and A. Ghazisaeidi, “Tx-Rx mode mismatch effects in gaussian-modulated CV-QKD”, in ECOC 2024; 50th Eu-ropean Conference on Optical Communication, 2024, pp. 1366–1369

    work page 2024

  9. [9]

    Contin-uous-variable quantum key distribution with Gaussian modulation: the theory of practical implementations,

    F. Laudenbach, C. Pacher, C.-H. F. Fung, et al., "Contin-uous-variable quantum key distribution with Gaussian modulation: the theory of practical implementations," Ad-vanced Quantum Technologies, vol. 1, 2018, Art. no. 1800011

    work page 2018

  10. [10]

    Pirandola, U

    S. Pirandola, U. L. Andersen, L. Banchi, et al., "Advances in quantum cryptography," Advances in Optics and Pho-tonics, vol. 12, no. 4, 2020, pp. 1012–1236. DOI: 10.1364/AOP.361502

    work page doi:10.1364/aop.361502 2020

  11. [11]

    Demonstration of probabilistic constellation shaping for continuous variable quantum key distribution,

    F. Roumestan, A. Ghazisaeidi, et al., “Demonstration of probabilistic constellation shaping for continuous variable quantum key distribution,” 2021 Optical Fiber Communi-cations Conference and Exhibition (OFC),

    work page 2021