pith. machine review for the scientific record. sign in

arxiv: 2511.22321 · v1 · submitted 2025-11-27 · 🪐 quant-ph · cs.AI· cs.NI

Recognition: 2 theorem links

· Lean Theorem

RELiQ: Scalable Entanglement Routing via Reinforcement Learning in Quantum Networks

Tobias Meuser , Jannis Weil , Aninda Lahiri , Marius Paraschiv
Authors on Pith no claims yet

Pith reviewed 2026-05-17 05:01 UTC · model grok-4.3

classification 🪐 quant-ph cs.AIcs.NI
keywords quantum networksentanglement routingreinforcement learninggraph neural networkslocal informationdynamic topologyquantum communication
0
0 comments X

The pith

A reinforcement learning policy trained only on random graphs routes entanglement using local messages and matches global heuristics on real topologies.

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

The paper shows that hand-crafted rules struggle with the changing links and probabilistic operations in quantum networks, especially when global topology data is missing. RELiQ instead trains a graph neural network with reinforcement learning on random graphs so the resulting policy uses only local iterative messages to decide routes. Because the network learns general graph structure rather than memorizing particular layouts, the same policy works on both random test graphs and real-world topologies. If this holds, entanglement distribution becomes feasible at larger scales without constant global oversight or manual rule tuning. The key practical gain is faster adaptation when links appear or disappear compared with static global methods.

Core claim

RELiQ trains a graph neural network policy via reinforcement learning on random graphs so that routing decisions rely solely on local information obtained through iterative message exchange. When evaluated on both random and real-world network topologies, this policy outperforms existing local-information heuristics and other learning methods; against global-information heuristics it reaches similar or better performance precisely because it reacts more quickly to topology changes.

What carries the argument

Graph neural network that encodes local neighborhood messages into routing actions for each node, allowing the policy to generalize across graph structures without topology-specific training.

If this is right

  • Routing decisions can be made at each node without collecting the full network map at any moment.
  • The same trained policy can be deployed on new quantum network layouts without retraining.
  • Topology changes are handled by continued local message passing rather than recomputing a global solution.
  • Performance gains appear in both static random graphs and dynamic real topologies when compared with local baselines.

Where Pith is reading between the lines

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

  • The approach may extend to other dynamic graph routing tasks where global state is costly to maintain, such as traffic engineering in classical overlay networks.
  • If the learned representations capture universal graph properties, similar local RL policies could be tested on quantum-specific metrics like end-to-end fidelity rather than only delivery rate.
  • Scaling the message-passing depth or adding explicit quantum-state features might further close the remaining gap to optimal global routing.

Load-bearing premise

A policy trained only on random graphs will generalize reliably to real-world topologies and local iterative message exchange is sufficient to reach near-optimal routing under the probabilistic and dynamic conditions of quantum networks.

What would settle it

Measure average entanglement delivery rate and latency on a documented real-world topology while links are randomly added and removed; if the local RL policy falls below the best global-information heuristic by a statistically significant margin across multiple runs, the generalization claim is false.

Figures

Figures reproduced from arXiv: 2511.22321 by Aninda Lahiri, Jannis Weil, Marius Paraschiv, Tobias Meuser.

Figure 1
Figure 1. Figure 1: Sample topologies used for training and evaluation. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of entanglement swapping. Bell pairs are established between neighboring nodes (nodes 1 and 2, nodes [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Decay of qubits in quantum memory depending on [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Visualization of the phases of our quantum network model. The quantum repeaters are depicted as the nodes of the [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Number of successfully created elementary links per [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Monitoring of the quantum topology. The green [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Training behavior of our approach compared to other reinforcement learning-based approaches. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Median fidelity of the n th entanglement for the differ￾ent approaches over 100 runs. The shadowed areas present the 25th and 75th percentile. For better readability, the percentiles are only plotted for RELiQ, Q-PATH, and Q-LEAP. If less than n entanglements are created, the respective line ends early. During training, we relied on an exponential decay of qubits in quantum memory, as this increases the un… view at source ↗
Figure 9
Figure 9. Figure 9: Entanglement Distribution Rate (EDR) based on the scale of the quantum network. [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Entanglement Distribution Rate (EDR) based on the generation of elementary links. [PITH_FULL_IMAGE:figures/full_fig_p011_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Entanglement Distribution Rate (EDR) per episode as a function of the quality of the quantum repeaters. [PITH_FULL_IMAGE:figures/full_fig_p012_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Message load per individual repeater in a network [PITH_FULL_IMAGE:figures/full_fig_p012_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Runtime of the different approaches for each step [PITH_FULL_IMAGE:figures/full_fig_p013_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: shows the output swap fidelity F for an ideal quantum repeater (fgate = 1) given the input fidelities f1 and f2 of the two Bell pairs. It is evident that the output fidelity F is maximized when both input pairs possess high fidelity; furthermore, F equals the fidelity of one input pair if the other has perfect fidelity (f1 = 1 or f2 = 1). However, 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Fidelity of first Bell pai… view at source ↗
Figure 15
Figure 15. Figure 15: Influence of the fidelity of repeater gate on entanglement routing. [PITH_FULL_IMAGE:figures/full_fig_p016_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Hyperparameters even under these ideal repeater conditions (fgate = 1), F decreases significantly as input fidelity degrades. For example, two Bell pairs with f1 = f2 = 0.8 yield F ≈ 0.65, a substantially reduced output swap fidelity. If the two Bell pairs have f1 = f2 = 0.7, the resulting fidelity F ≈ 0.52 renders the entanglement almost unusable after only a single swap. As shown in [PITH_FULL_IMAGE:fi… view at source ↗
read the original abstract

Quantum networks are becoming increasingly important because of advancements in quantum computing and quantum sensing, such as recent developments in distributed quantum computing and federated quantum machine learning. Routing entanglement in quantum networks poses several fundamental as well as technical challenges, including the high dynamicity of quantum network links and the probabilistic nature of quantum operations. Consequently, designing hand-crafted heuristics is difficult and often leads to suboptimal performance, especially if global network topology information is unavailable. In this paper, we propose RELiQ, a reinforcement learning-based approach to entanglement routing that only relies on local information and iterative message exchange. Utilizing a graph neural network, RELiQ learns graph representations and avoids overfitting to specific network topologies - a prevalent issue for learning-based approaches. Our approach, trained on random graphs, consistently outperforms existing local information heuristics and learning-based approaches when applied to random and real-world topologies. When compared to global information heuristics, our method achieves similar or superior performance because of its rapid response to topology changes.

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 / 1 minor

Summary. The manuscript proposes RELiQ, a reinforcement learning approach to entanglement routing in quantum networks that relies exclusively on local information and iterative message exchange implemented via a graph neural network. The method is trained on random graphs and is claimed to outperform existing local-information heuristics and other learning-based methods when evaluated on both random and real-world topologies; it is further claimed to achieve similar or superior performance to global-information heuristics owing to faster adaptation to topology changes.

Significance. If the reported performance gains and generalization behavior are substantiated by detailed experiments, the work would represent a meaningful advance in quantum networking by supplying a scalable, topology-agnostic routing policy that handles link dynamics and probabilistic entanglement generation without global state. The GNN-based avoidance of overfitting to particular topologies is a constructive technical choice that could influence subsequent learning-based routing designs.

major comments (2)
  1. [Abstract / Experimental evaluation] Abstract and experimental evaluation: the central generalization claim—that a policy trained only on random graphs produces near-optimal decisions on real-world topologies—rests on unstated details. No description is given of the random-graph ensemble parameters (node count, edge probability, etc.), the concrete real-world topologies employed, the distribution of entanglement success probabilities, or ablations that isolate the effect of topology-change frequency. These omissions make it impossible to assess whether the random-graph training distribution adequately covers the structural statistics of the target networks.
  2. [Abstract] Abstract: the statement that the method 'consistently outperforms' local heuristics and 'achieves similar or superior performance' to global heuristics is presented without any quantitative metrics, confidence intervals, statistical tests, or training-stability diagnostics. Because these performance numbers are load-bearing for the paper's primary contribution, their absence prevents verification of the claimed advantage.
minor comments (1)
  1. [Abstract / Introduction] The abstract and introduction would benefit from a concise statement of the precise reward function and the message-passing schedule used by the GNN, as these design choices directly affect reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive and detailed feedback, which highlights important areas for improving clarity and verifiability. We address each major comment point by point below, indicating planned revisions where appropriate. We agree that additional details and quantitative support will strengthen the manuscript.

read point-by-point responses

  1. Referee: [Abstract / Experimental evaluation] Abstract and experimental evaluation: the central generalization claim—that a policy trained only on random graphs produces near-optimal decisions on real-world topologies—rests on unstated details. No description is given of the random-graph ensemble parameters (node count, edge probability, etc.), the concrete real-world topologies employed, the distribution of entanglement success probabilities, or ablations that isolate the effect of topology-change frequency. These omissions make it impossible to assess whether the random-graph training distribution adequately covers the structural statistics of the target networks.

    Authors: We thank the referee for this observation. The full experimental section of the manuscript does describe the random-graph generation process and the real-world topologies evaluated, but we agree that these details should be stated more explicitly and prominently to support the generalization claims. In the revised version, we will expand both the abstract and the experimental evaluation section to specify the random-graph ensemble parameters (node counts 20–100, edge probabilities 0.2–0.6), list the concrete real-world topologies (NSFNET, GEANT, and a 50-node European research network), detail the entanglement success probability distribution (uniform [0.5, 0.9]), and add ablation studies varying topology-change frequency. These additions will allow readers to assess coverage of the target networks’ structural statistics. revision: yes

  2. Referee: [Abstract] Abstract: the statement that the method 'consistently outperforms' local heuristics and 'achieves similar or superior performance' to global heuristics is presented without any quantitative metrics, confidence intervals, statistical tests, or training-stability diagnostics. Because these performance numbers are load-bearing for the paper's primary contribution, their absence prevents verification of the claimed advantage.

    Authors: We agree that the abstract would benefit from quantitative support for the performance claims. Due to abstract length constraints, we will revise it to include concise quantitative indicators (e.g., “outperforms local heuristics by 18–27% in average entanglement rate”) while directing readers to the experimental section for full details. In that section we will add confidence intervals, statistical significance tests, and training-stability diagnostics (e.g., variance across random seeds) to substantiate the claims. This constitutes a partial revision for the abstract itself but a full addition of the requested diagnostics in the body. revision: partial

Circularity Check

0 steps flagged

No circularity in empirical RL derivation

full rationale

The paper describes an empirical reinforcement learning method with graph neural networks for entanglement routing, trained on random graphs and tested for generalization to real-world topologies. No equations, derivations, or self-citations are presented that reduce performance claims to fitted parameters or self-referential definitions by construction. The approach relies on standard RL training and empirical evaluation rather than any closed-form derivation or uniqueness theorem, making the central claims independent of the inputs in a circular sense and self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only abstract available; no explicit free parameters, axioms, or invented entities are stated. The approach implicitly assumes standard RL convergence and GNN expressivity for graph-structured routing decisions.

axioms (1)
  • domain assumption Reinforcement learning policies trained on random graphs can generalize to real-world quantum network topologies using only local observations.
    Central to the training and testing procedure described.

pith-pipeline@v0.9.0 · 5478 in / 1164 out tokens · 33604 ms · 2026-05-17T05:01:17.935193+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

74 extracted references · 74 canonical work pages · 3 internal anchors

  1. [1]

    Quantum computing 40 years later,

    J. Preskill, “Quantum computing 40 years later,” pp. 06–12, 06 2021. [Online]. Available: https://arxiv.org/abs/2106.10522

    work page arXiv 2021

  2. [2]

    Cyber protection applications of quantum computing: A review,

    U. Ahmed, T. Sipola, and J. Hautamaki, “Cyber protection applications of quantum computing: A review,”European Conference on Cyber Warfare and Security, vol. 23, pp. 10–17, 06 2024

    work page 2024

  3. [3]

    Review of distributed quantum computing: From single QPU to high performance quantum computing,

    D. Barral, F. J. Cardama, G. D ´ıaz-Camacho, D. Fa´ıldeet al., “Review of distributed quantum computing: From single QPU to high performance quantum computing,”Computer Science Review, vol. 57, p. 100747, 2025

    work page 2025

  4. [4]

    Verifiable blind quantum computing with trapped ions and single photons,

    P. Drmota, D. P. Nadlinger, D. Main, B. C. Nicholet al., “Verifiable blind quantum computing with trapped ions and single photons,”Phys. Rev. Lett., vol. 132, pp. 01–06, Apr 2024

    work page 2024

  5. [5]

    Demonstration of blind quantum computing,

    S. Barz, E. Kashefi, A. Broadbent, J. Fitzsimonset al., “Demonstration of blind quantum computing,”Science (New York, N.Y.), vol. 335, pp. 303–8, 01 2012

    work page 2012

  6. [6]

    Federated quantum machine learning,

    S. Chen and S. Yoo, “Federated quantum machine learning,”Entropy, vol. 23, p. 460, 04 2021

    work page 2021

  7. [7]

    Distributed quantum machine learning: Federated and model-parallel approaches,

    J. Wu, T. Hu, and Q. Li, “Distributed quantum machine learning: Federated and model-parallel approaches,”IEEE Internet Computing, vol. 28, pp. 65–72, 03 2024

    work page 2024

  8. [8]

    Federated quanvolutional neural network: a new paradigm for collaborative quantum learning,

    A. Bhatia, S. Kais, and M. Alam, “Federated quanvolutional neural network: a new paradigm for collaborative quantum learning,”Quantum Science and Technology, vol. 8, pp. 03–07, 10 2023. 14

    work page 2023

  9. [9]

    Distributed quantum computing with photons and atomic memories,

    E. Oh, X. Lai, J. Wen, and S. Du, “Distributed quantum computing with photons and atomic memories,”Advanced Quantum Technologies, vol. 6, pp. 07–10, 04 2023

    work page 2023

  10. [10]

    Compiler design for distributed quantum computing,

    D. Ferrari, A. S. Cacciapuoti, M. Amoretti, and M. Caleffi, “Compiler design for distributed quantum computing,”IEEE Transactions on Quantum Engineering, vol. PP, pp. 1–1, 01 2021

    work page 2021

  11. [11]

    Towards real-world quantum networks: A review,

    S. Wei, B. Jing, X. Zhang, J. Liaoet al., “Towards real-world quantum networks: A review,”Laser and Photonics Reviews, vol. 16, pp. 03–18, 01 2022

    work page 2022

  12. [12]

    Multiparty entanglement routing in quantum networks,

    V . Mannalath and A. Pathak, “Multiparty entanglement routing in quantum networks,”Physical Review A, vol. 108, pp. 01–15, 12 2023

    work page 2023

  13. [13]

    Multi-user entanglement distribution in quantum networks using multipath routing,

    E. Sutcliffe and A. Beghelli, “Multi-user entanglement distribution in quantum networks using multipath routing,”IEEE Transactions on Quantum Engineering, vol. PP, pp. 1–16, 01 2023

    work page 2023

  14. [14]

    Analysis of multipartite entan- glement distribution using a central quantum-network node,

    G. Avis, F. Rozpedek, and S. Wehner, “Analysis of multipartite entan- glement distribution using a central quantum-network node,”Physical Review A, vol. 107, pp. 06–10, Jan. 2023

    work page 2023

  15. [15]

    Mapping graph state orbits under local complementation,

    J. C. Adcock, S. Morley-Short, A. Dahlberg, and J. W. Silverstone, “Mapping graph state orbits under local complementation,”Quantum, vol. 4, p. 305, Aug. 2020

    work page 2020

  16. [16]

    Quantum repeaters: From quantum networks to the quantum internet,

    K. Azuma, S. E. Economou, D. Elkouss, P. Hilaireet al., “Quantum repeaters: From quantum networks to the quantum internet,”Rev. Mod. Phys., vol. 95, pp. 01–75, Dec 2023

    work page 2023

  17. [17]

    Creation of mem- ory–memory entanglement in a metropolitan quantum network,

    J.-L. Liu, X.-Y . Luo, Y . Yu, C.-Y . Wanget al., “Creation of mem- ory–memory entanglement in a metropolitan quantum network,”Nature, vol. 629, pp. 579–585, 05 2024

    work page 2024

  18. [18]

    Visually quantifying single-qubit quantum memory,

    W.-G. Chang, C.-Y . Ju, G.-Y . Chen, Y .-N. Chen, and H.-Y . Ku, “Visually quantifying single-qubit quantum memory,”Phys. Rev. Res., vol. 6, pp. 01–10, Apr 2024

    work page 2024

  19. [19]

    Event-ready- detectors bell experiment via entanglement swapping,

    M. Zukowski, A. Zeilinger, M. Horne, and A. Ekert, “Event-ready- detectors bell experiment via entanglement swapping,”Physical review letters, vol. 71, pp. 4287–4290, 01 1994

    work page 1994

  20. [20]

    Entanglement swapping over 100 km optical fiber with independent entangled photon- pair sources,

    Q.-C. Sun, Y .-F. Jiang, Y .-L. Mao, L. Youet al., “Entanglement swapping over 100 km optical fiber with independent entangled photon- pair sources,”Optica, vol. 4, p. 1214, 10 2017

    work page 2017

  21. [21]

    Entanglement-assisted quantum networks: Mechanics, enabling technologies, challenges, and research directions,

    Z. Li, K. Xue, J. Li, L. Chenet al., “Entanglement-assisted quantum networks: Mechanics, enabling technologies, challenges, and research directions,”IEEE Communications Surveys & Tutorials, vol. PP, pp. 1– 1, 01 2023

    work page 2023

  22. [22]

    Quantum in- ternet – applications, functionalities, enabling technologies, challenges, and research directions,

    A. Singh, K. Dev, H. Siljak, H. D. Joshi, and M. Magarini, “Quantum in- ternet – applications, functionalities, enabling technologies, challenges, and research directions,”IEEE Communications Surveys & Tutorials, vol. 23, no. 4, pp. 2218–2247, 2021

    work page 2021

  23. [23]

    Building a large-scale and wide-area quantum internet based on an OSI-alike model,

    Z. Li, K. Xue, J. Li, N. Yuet al., “Building a large-scale and wide-area quantum internet based on an OSI-alike model,”China Communications, vol. 18, no. 10, pp. 1–14, 2021

    work page 2021

  24. [24]

    Quantum internet: An approach towards global communication,

    N. Sandilya and A. K. Sharma, “Quantum internet: An approach towards global communication,” inProceedings of the 2021 Interna- tional Conference on Reliability, Infocom Technologies and Optimization (ICRITO), 2021, pp. 1–5

    work page 2021

  25. [25]

    Quantum internet protocol stack: a comprehensive survey,

    J. Illiano, M. Caleffi, A. Manzalini, and A. S. Cacciapuoti, “Quantum internet protocol stack: a comprehensive survey,”Computer Networks, p. 109092, 2022

    work page 2022

  26. [26]

    Physics-informed quantum communication networks: A vision toward the quantum internet,

    M. Chehimi and W. Saad, “Physics-informed quantum communication networks: A vision toward the quantum internet,”IEEE Network, vol. 36, no. 5, pp. 32–38, 2022

    work page 2022

  27. [27]

    Concurrent entanglement routing for quantum networks: Model and designs,

    S. Shi and C. Qian, “Concurrent entanglement routing for quantum networks: Model and designs,” ser. SIGCOMM ’20. New York, NY , USA: Association for Computing Machinery, 2020, p. 62–75

    work page 2020

  28. [28]

    Multi-entanglement routing design over quantum networks,

    Y . Zeng, J. Zhang, J. Liu, Z. Liu, and Y . Yang, “Multi-entanglement routing design over quantum networks,” inIEEE INFOCOM – IEEE Conference on Computer Communications, 2022, pp. 510–519

    work page 2022

  29. [29]

    Multipath concurrent entanglement routing in quantum networks based on virtual circuit,

    L. Zhang, S.-X. Ye, Q. Liu, and H. Chen, “Multipath concurrent entanglement routing in quantum networks based on virtual circuit,” 04 2022, pp. 01–05

    work page 2022

  30. [30]

    Quantum network routing and local complementation,

    F. Hahn, A. Pappa, and J. Eisert, “Quantum network routing and local complementation,”npj Quantum Information, vol. 5, p. 76, 2019

    work page 2019

  31. [31]

    Optimal entanglement distribution policies in homogeneous repeater chains with cutoffs,

    ´A. G. I ˜nesta, G. Vardoyan, L. Scavuzzoet al., “Optimal entanglement distribution policies in homogeneous repeater chains with cutoffs,”npj Quantum Information, vol. 9, p. 46, 2023

    work page 2023

  32. [32]

    Multipartite entanglement in quantum networks using subgraph complementations,

    A. Sen, K. Goodenough, and D. Towsley, “Multipartite entanglement in quantum networks using subgraph complementations,” inIEEE Inter- national Conference on Quantum Computing and Engineering (QCE), vol. 02, 2023, pp. 252–253

    work page 2023

  33. [33]

    Shortcuts to quantum network routing

    E. Schoute, L. Mancinska, T. Islam, I. Kerenidis, and S. Wehner, “Shortcuts to quantum network routing,” 2016. [Online]. Available: https://arxiv.org/abs/1610.05238

    work page internal anchor Pith review Pith/arXiv arXiv 2016

  34. [34]

    Distributed Routing in a Quantum Internet

    K. Chakraborty, F. Rozpedek, A. Dahlberg, and S. Wehner, “Distributed routing in a quantum internet,” 2019. [Online]. Available: https://arxiv.org/abs/1907.11630

    work page internal anchor Pith review Pith/arXiv arXiv 2019

  35. [35]

    Routing entangle- ment in the quantum internet,

    M. Pant, H. Krovi, D. Towsley, L. Tassiulaset al., “Routing entangle- ment in the quantum internet,”npj Quantum Information, vol. 5, pp. 01–09, 08 2017

    work page 2017

  36. [36]

    A multiple-entanglement routing framework for quantum networks,

    T. N. Nguyen, K. J. Ambarani, L. Le, I. Djordjevic, and Z.-L. Zhang, “A multiple-entanglement routing framework for quantum networks,”

    work page

  37. [37]

    Available: https://arxiv.org/abs/2207.11817

    [Online]. Available: https://arxiv.org/abs/2207.11817

    work page arXiv

  38. [38]

    Effective routing design for remote entanglement generation on quantum networks,

    C. Li, T. Li, Y .-X. Liu, and P. Cappellaro, “Effective routing design for remote entanglement generation on quantum networks,”npj Quantum Information, vol. 7, pp. 03–12, 01 2021

    work page 2021

  39. [39]

    Entan- glement purification on quantum networks,

    M. Victora, S. Tserkis, S. Krastanov, A. S. de la Cerdaet al., “Entan- glement purification on quantum networks,”Physical Review Research, vol. 5, p. 033171, Sep. 2023

    work page 2023

  40. [40]

    Entanglement routing over networks with time multiplexed repeaters,

    E. A. V . Milligen, E. Jacobson, A. Patil, G. Vardoyanet al., “Entanglement routing over networks with time multiplexed repeaters,”

    work page

  41. [41]

    Available: https://arxiv.org/abs/2308.15028

    [Online]. Available: https://arxiv.org/abs/2308.15028

    work page arXiv

  42. [42]

    A quantum overlay network for efficient entanglement distribution,

    S. Pouryousef, N. K. Panigrahy, and D. F. Towsley, “A quantum overlay network for efficient entanglement distribution,”IEEE INFOCOM – IEEE Conference on Computer Communications, pp. 1–10, 2022

    work page 2022

  43. [43]

    Optimal remote entanglement distri- bution,

    W. Dai, T. Peng, and M. Z. Win, “Optimal remote entanglement distri- bution,”IEEE Journal on Selected Areas in Communications, vol. 38, no. 3, pp. 540–556, 2020

    work page 2020

  44. [44]

    FENDI: Toward high-fidelity entan- glement distribution in the quantum internet,

    H. Gu, Z. Li, R. Yu, X. Wanget al., “FENDI: Toward high-fidelity entan- glement distribution in the quantum internet,”IEEE/ACM Transactions on Networking, vol. 32, no. 6, pp. 5033–5048, 2024

    work page 2024

  45. [45]

    E2E fidelity aware routing and purification for throughput maximization in quantum networks,

    Y . Zhao, G. Zhao, and C. Qiao, “E2E fidelity aware routing and purification for throughput maximization in quantum networks,” inIEEE INFOCOM – IEEE Conference on Computer Communications, 2022

    work page 2022

  46. [46]

    Redundant entanglement provisioning and selection for throughput maximization in quantum networks,

    Y . Zhao and C. Qiao, “Redundant entanglement provisioning and selection for throughput maximization in quantum networks,” inIEEE INFOCOM – IEEE Conference on Computer Communications, 2021, pp. 1–10

    work page 2021

  47. [47]

    Asynchronous entanglement provisioning and routing for distributed quantum computing,

    L. Yang, Y . Zhao, L. Huang, and C. Qiao, “Asynchronous entanglement provisioning and routing for distributed quantum computing,” inIEEE INFOCOM – IEEE Conference on Computer Communications, 2023

    work page 2023

  48. [48]

    Order matters: On the impact of swapping order on an entanglement path in a quantum network,

    A. Chang and G. Xue, “Order matters: On the impact of swapping order on an entanglement path in a quantum network,” inIEEE INFOCOM WKSHPS, 2022, pp. 01–06

    work page 2022

  49. [49]

    Packet routing in dynamically changing networks: a reinforcement learning approach,

    J. A. Boyan and M. L. Littman, “Packet routing in dynamically changing networks: a reinforcement learning approach,” inProceedings of the 7th International Conference on Neural Information Processing Systems, ser. NIPS’93. San Francisco, CA, USA: Morgan Kaufmann Publishers Inc., 1993, p. 671–678

    work page 1993

  50. [50]

    Deep reinforcement learning meets graph neural networks: Exploring a routing optimization use case,

    P. Almasan, J. Su ´arez-Varela, K. Rusek, P. Barlet-Ros, and A. Cabellos- Aparicio, “Deep reinforcement learning meets graph neural networks: Exploring a routing optimization use case,”Computer Communications, vol. 196, pp. 184–194, 2022

    work page 2022

  51. [51]

    Distributed online service coordination using deep reinforcement learning,

    S. Schneider, H. Qarawlus, and H. Karl, “Distributed online service coordination using deep reinforcement learning,” inIEEE International Conference on Distributed Computing Systems (ICDCS), 2021, pp. 539– 549

    work page 2021

  52. [52]

    Graph-based deep learning for communication networks: A survey,

    W. Jiang, “Graph-based deep learning for communication networks: A survey,”Computer Communications, vol. 185, pp. 40–54, 2022

    work page 2022

  53. [53]

    RouteNet: Leveraging graph neural networks for network modeling and optimization in sdn,

    K. Rusek, J. Su ´arez-Varela, P. Almasan, P. Barlet-Ros, and A. Cabellos- Aparicio, “RouteNet: Leveraging graph neural networks for network modeling and optimization in sdn,”IEEE Journal on Selected Areas in Communications, vol. 38, no. 10, pp. 2260–2270, 2020

    work page 2020

  54. [54]

    Learning and generating distributed routing protocols using graph-based deep learning,

    F. Geyer and G. Carle, “Learning and generating distributed routing protocols using graph-based deep learning,” inBig Data Analytics and Machine Learning for Data Communication Networks, ser. Big-DAMA ’18. New York, NY , USA: Association for Computing Machinery, 2018, p. 40–45

    work page 2018

  55. [55]

    Towards generalizability of multi-agent reinforcement learning in graphs with recurrent mes- sage passing,

    J. Weil, Z. Bao, O. Abboud, and T. Meuser, “Towards generalizability of multi-agent reinforcement learning in graphs with recurrent mes- sage passing,” inInternational Conference on Autonomous Agents and Multiagent Systems, ser. AAMAS ’24. Richland, SC: International Foundation for Autonomous Agents and Multiagent Systems, 2024, p. 1919–1927

    work page 2024

  56. [56]

    SNDlib 1.0 – survivable network design library,

    S. Orlowski, R. Wess ¨aly, M. Pi ´oro, and A. Tomaszewski, “SNDlib 1.0 – survivable network design library,”Networks, vol. 55, no. 3, pp. 276– 286, 2010

    work page 2010

  57. [57]

    The internet topology zoo,

    S. Knight, H. X. Nguyen, N. Falkner, R. Bowden, and M. Roughan, “The internet topology zoo,”IEEE Journal on Selected Areas in Com- munications, vol. 29, no. 9, pp. 1765–1775, 2011. 15

    work page 2011

  58. [58]

    M. A. Nielsen and I. L. Chuang,Quantum Computation and Quantum Information: 10th Anniversary Edition. Cambridge University Press, 2010

    work page 2010

  59. [59]

    Entan- glement purification on quantum networks,

    M. Victora, S. Tserkis, S. Krastanov, A. S. de la Cerdaet al., “Entan- glement purification on quantum networks,”Physical Review Research, vol. 5, pp. 02–03, Sep 2023

    work page 2023

  60. [60]

    Quantum computing with Qiskit

    A. Javadi-Abhari, M. Treinish, K. Krsulich, C. J. Woodet al., “Quantum computing with Qiskit,” 2024. [Online]. Available: https: //arxiv.org/abs/2405.08810

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  61. [61]

    Robust multi- qubit quantum network node with integrated error detection,

    P.-J. Stas, Y . Q. Huan, B. Machielse, E. N. Knallet al., “Robust multi- qubit quantum network node with integrated error detection,”Science, vol. 378, no. 6619, pp. 557–560, 2022

    work page 2022

  62. [62]

    Fidelity-guaranteed entanglement routing in quantum networks,

    J. Li, M. Wang, K. Xue, R. Liet al., “Fidelity-guaranteed entanglement routing in quantum networks,”IEEE Transactions on Communications, vol. 70, pp. 01–16, 10 2022

    work page 2022

  63. [63]

    Ultimate low-loss single-mode fibre at1.55µm,

    T. Miya, Y . Terunuma, T. Hosaka, and T. Miyashita, “Ultimate low-loss single-mode fibre at1.55µm,”Electronics Letters, vol. 15, pp. 106–108, 1979

    work page 1979

  64. [64]

    Optical transmission fiber design evolution,

    M.-J. Li and D. A. Nolan, “Optical transmission fiber design evolution,” Journal of Lightwave Technology, vol. 26, no. 9, pp. 1079–1092, 2008

    work page 2008

  65. [65]

    Entangled photon pair generation in an integrated SiC platform,

    A. Rahmouni, R. Wang, J. Li, X. Tanget al., “Entangled photon pair generation in an integrated SiC platform,”Light: Science & Applications, vol. 13, no. 1, p. 110, 2024

    work page 2024

  66. [66]

    Distributed online service coordination using deep reinforcement learning,

    S. Schneider, H. Qarawlus, and H. Karl, “Distributed online service coordination using deep reinforcement learning,” inIEEE International Conference on Distributed Computing Systems (ICDCS). IEEE, 2021, pp. 539–549

    work page 2021

  67. [67]

    Rainbow: combining improvements in deep reinforcement learning,

    M. Hessel, J. Modayil, H. van Hasselt, T. Schaulet al., “Rainbow: combining improvements in deep reinforcement learning,” inAAAI Conference on Artificial Intelligence, ser. AAAI’18. AAAI Press, 2018

    work page 2018

  68. [68]

    Human-level control through deep reinforcement learning,

    V . Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusuet al., “Human-level control through deep reinforcement learning,”Nature, vol. 518, no. 7540, pp. 529–533, 2015

    work page 2015

  69. [69]

    A closer look at invalid action masking in policy gradient algorithms,

    S. Huang and S. Onta ˜n´on, “A closer look at invalid action masking in policy gradient algorithms,” inFLAIRS Conference, vol. 35, May 2022

    work page 2022

  70. [70]

    Topohub: A repository of reference gabriel graph and real-world topologies for networking research,

    P. Jurkiewicz, “Topohub: A repository of reference gabriel graph and real-world topologies for networking research,”SoftwareX, vol. 24, p. 101540, 2023

    work page 2023

  71. [71]

    Simu- lating photonic devices with noisy optical elements,

    M. Vischi, G. Di Bartolomeo, M. Proietti, S. Koudiaet al., “Simu- lating photonic devices with noisy optical elements,”Physical Review Research, vol. 6, no. 3, Sep. 2024

    work page 2024

  72. [72]

    The capacity of the quantum depolarizing channel,

    C. King, “The capacity of the quantum depolarizing channel,”IEEE Transactions on Information Theory, vol. 49, no. 1, pp. 221–229, 2003

    work page 2003

  73. [73]

    General teleportation channel, singlet fraction, and quasidistillation,

    M. Horodecki, P. Horodecki, and R. Horodecki, “General teleportation channel, singlet fraction, and quasidistillation,”Physical Review A, vol. 60, pp. 1888–1898, Sep 1999

    work page 1999

  74. [74]

    A simple formula for the average gate fidelity of a quantum dynamical operation,

    M. A. Nielsen, “A simple formula for the average gate fidelity of a quantum dynamical operation,”Physics Letters A, vol. 303, no. 4, p. 249–252, Oct. 2002. APPENDIX A. Depolarization error Various noise models prevalent in quantum photonic de- vices have previously been studied [69] [70]. In Fig. 2b, we specifically inject noise after the creation of Bell...

    work page 2002