arxiv: 2605.12365 · v1 · submitted 2026-05-12 · 🪐 quant-ph · cs.AI

Recognition: no theorem link

QAP-Router: Tackling Qubit Routing as Dynamic Quadratic Assignment with Reinforcement Learning

Ankit Kulshrestha, Ilya Safro, Kien X. Nguyen, Xiaoyuan Liu

Pith reviewed 2026-05-13 04:41 UTC · model grok-4.3

classification 🪐 quant-ph cs.AI

keywords qubit routingquadratic assignment problemreinforcement learningquantum compilationtransformer attentionCNOT reductionlookahead policy

0 comments

The pith

Framing qubit routing as a dynamic quadratic assignment problem lets a reinforcement learning policy with lookahead and transformer attention cut added CNOT gates by 12 to 30 percent.

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

Qubit routing requires inserting swaps so that logical qubits can interact on fixed hardware, but each swap adds CNOT gates that increase circuit depth and error. The paper claims this process can be cast as a dynamic quadratic assignment problem in which gate interactions act as a flow matrix and hardware links act as a distance matrix; the product of the two matrices then supplies a single scalar reward for reinforcement learning. A solution-aware transformer policy encodes the current flow-distance coupling inside its attention layers, while a lookahead step blends future assignments into the same objective. On 1831 circuits drawn from three standard suites the resulting routings require 15.7 percent, 30.4 percent, and 12.1 percent fewer CNOTs than industry compilers. If the claim holds, compilers could deliver shallower circuits that run on today’s noisy hardware before decoherence dominates.

Core claim

The central claim is that qubit routing is best solved by maintaining a dynamic quadratic assignment objective whose reward equals the sum over logical gate flows times physical distances; this objective is optimized by a reinforcement-learning policy whose transformer backbone injects the flow-distance product into attention and whose lookahead mechanism evaluates candidate assignments several steps ahead. The formulation therefore unifies local gate placement with global topology cost inside a single learned policy rather than relying on hand-crafted heuristics or generic sequential decision models.

What carries the argument

The dynamic quadratic assignment objective that treats logical gate interactions as a flow matrix and hardware connectivity as a distance matrix, used both to define the reinforcement-learning reward and to shape the attention inside a solution-aware transformer policy equipped with lookahead.

If this is right

Local routing decisions become globally consistent because each step optimizes the same flow-distance product rather than a myopic heuristic.
The transformer attention layers directly exploit the QAP structure, allowing the policy to weigh future gate interactions against current hardware distances.
Lookahead integration reduces error accumulation that arises when early swaps lock the circuit into costly later configurations.
The measured CNOT reductions hold across real-world circuits from multiple independent suites, suggesting the method is not tuned to a single benchmark family.

Where Pith is reading between the lines

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

The same flow-distance framing could be reused for other compilation subtasks such as initial placement or gate scheduling by redefining the flow matrix accordingly.
On larger devices the learned policy might serve as a fast heuristic inside hybrid classical-quantum compilers that alternate between exact solvers and the RL router.
Because the reward is explicitly a matrix product, the approach could be extended to weighted costs that incorporate measured gate error rates or decoherence times without changing the policy architecture.

Load-bearing premise

The dynamic QAP objective together with the solution-aware transformer policy and lookahead will continue to yield globally efficient routings on circuits and hardware topologies outside the three benchmark suites tested.

What would settle it

A head-to-head run on a fresh collection of circuits drawn from an application domain or device size not represented in MQTBench, AgentQ, or QUEKO, in which the method produces no reduction in CNOT count relative to the same industry baselines, would falsify the generality of the performance claim.

Figures

Figures reproduced from arXiv: 2605.12365 by Ankit Kulshrestha, Ilya Safro, Kien X. Nguyen, Xiaoyuan Liu.

**Figure 2.** Figure 2: Motivation for integration of QAP objective to the reward function. At time slice [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Logical circuit to flow matrices conversion procedure and state space encoding pipeline. (a) [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

**Figure 4.** Figure 4: Comparison on the number of inserted CNOT gates across different look-ahead horizons [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗

**Figure 5.** Figure 5: 2D Grid device topology for 12, 16 and 20 qubits. [PITH_FULL_IMAGE:figures/full_fig_p015_5.png] view at source ↗

**Figure 6.** Figure 6: IBM Tokyo-like device topology for 12, 16 and 20 qubits. [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗

**Figure 7.** Figure 7: Comparison on the number of inserted CNOT gates between [PITH_FULL_IMAGE:figures/full_fig_p020_7.png] view at source ↗

**Figure 8.** Figure 8: Comparison on the number of inserted CNOT gates between [PITH_FULL_IMAGE:figures/full_fig_p021_8.png] view at source ↗

**Figure 9.** Figure 9: Comparison on the number of inserted CNOT gates between [PITH_FULL_IMAGE:figures/full_fig_p022_9.png] view at source ↗

**Figure 10.** Figure 10: Comparison on the number of inserted CNOT gates between [PITH_FULL_IMAGE:figures/full_fig_p022_10.png] view at source ↗

**Figure 11.** Figure 11: Comparison on the number of inserted CNOT gates between [PITH_FULL_IMAGE:figures/full_fig_p023_11.png] view at source ↗

**Figure 12.** Figure 12: Comparison on the number of inserted CNOT gates between [PITH_FULL_IMAGE:figures/full_fig_p023_12.png] view at source ↗

read the original abstract

Qubit routing is a fundamental problem in quantum compilation, known to be NP-hard. Its dynamic nature makes local routing decisions propagate and compound over time, making global efficient solutions challenging. Existing heuristic methods rely on local rules with limited lookahead, while recent learning-based approaches often treat routing as a generic sequential decision problem without fully exploiting its underlying structure. In this paper, we introduce QAP-Router, framing qubit routing based on a dynamic Quadratic Assignment Problem (QAP) formulation. By modeling logical interactions, or quantum gates, as flow matrices and hardware topology as a distance matrix, our approach captures the interaction-distance coupling in a unified objective, which defines the reward in the reinforcement learning environment. To further exploit this structure, the policy network employs a solution-aware Transformer backbone that encodes the interaction between the flow matrix and the distance matrix into the attention mechanism. We also integrate a lookahead mechanism that blends naturally into the QAP framework, preventing myopic decisions. Extensive experiments on 1,831 real-world quantum circuits from the MQTBench, AgentQ and QUEKO datasets show that our method substantially reduces the CNOT gate count of routed circuits by 15.7%, 30.4% and 12.1%, respectively, relative to existing industry compilers.

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.

Desk Editor's Note private letter to a colleague

QAP-Router casts qubit routing as dynamic QAP solved by RL with a solution-aware Transformer and lookahead, reporting CNOT reductions on three benchmark suites, but the evaluation leaves baseline details and generalization unaddressed.

read the letter

The paper's core move is to model qubit routing explicitly as a dynamic quadratic assignment problem: logical gate interactions become a flow matrix, hardware topology a distance matrix, and the combined objective becomes the RL reward. The policy then uses a Transformer whose attention is conditioned on the current solution state plus a lookahead step that fits naturally inside the QAP structure. This is a cleaner way to embed the combinatorial coupling than generic sequential RL formulations that ignore the flow-distance interaction until the reward is computed. The experiments run on 1,831 circuits across MQTBench, AgentQ, and QUEKO and claim CNOT reductions of 15.7%, 30.4%, and 12.1% versus industry compilers, which is the scale that would matter for near-term compilation if the numbers are robust. The modeling itself looks internally consistent and avoids the circularity trap of fitting parameters to the same metric being optimized. The main soft spots are in the empirical section. The abstract gives headline percentages without reporting variance across seeds, exact baseline compiler versions and settings, or scaling behavior with qubit count and circuit depth. Because the policy is learned, any mismatch in interaction-pattern statistics between the training distribution and new circuits or hardware graphs can shrink the reported edge, and the write-up does not describe cross-dataset hold-outs or topology-agnostic tests. Those gaps are fixable but currently make the strength of the advantage hard to judge. This is aimed at people already working on learning-based quantum compilers or on RL for structured combinatorial problems. A reading group could usefully dissect the attention mechanism and the lookahead integration. It deserves peer review because the formulation is explicit, the dataset size is reasonable, and the claimed gains are large enough to warrant closer inspection even if the evaluation needs tightening.

Referee Report

2 major / 1 minor

Summary. The paper introduces QAP-Router, framing qubit routing as a dynamic Quadratic Assignment Problem (QAP) solved via reinforcement learning. Logical gate interactions are modeled as flow matrices and hardware topology as a distance matrix to define the RL reward; the policy uses a solution-aware Transformer that incorporates flow-distance interactions into attention, augmented by a lookahead mechanism. Experiments on 1,831 circuits from MQTBench, AgentQ, and QUEKO report CNOT-count reductions of 15.7%, 30.4%, and 12.1% relative to existing industry compilers.

Significance. If the empirical margins hold under rigorous controls, the work is significant because it supplies a structured, globally-aware objective for an NP-hard dynamic routing task that existing local heuristics and generic RL pipelines do not fully exploit. Credit is due for the explicit QAP reward construction and the integration of the flow-distance coupling directly into the Transformer attention and lookahead, which are falsifiable design choices that can be tested on new topologies.

major comments (2)

[Abstract] Abstract: the headline performance claims (15.7%, 30.4%, 12.1% CNOT reduction) are presented without naming the exact baseline compilers, their versions or optimization flags, the number of independent runs, or any measure of variance or statistical significance; these omissions are load-bearing for the central empirical assertion.
[Experiments] Experiments section: no cross-dataset hold-out, no scaling curves with qubit count or depth, and no topology-agnostic test set are described, so it is impossible to determine whether the reported gains arise from the dynamic-QAP formulation or from statistical similarity between the training and test circuit ensembles.

minor comments (1)

[Method] The notation for the dynamic flow matrix update rule should be stated explicitly with an equation number so that the lookahead integration can be verified without ambiguity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. The comments highlight important aspects of clarity and experimental rigor that we will address. We respond point-by-point to the major comments below and indicate the revisions planned for the manuscript.

read point-by-point responses

Referee: [Abstract] Abstract: the headline performance claims (15.7%, 30.4%, 12.1% CNOT reduction) are presented without naming the exact baseline compilers, their versions or optimization flags, the number of independent runs, or any measure of variance or statistical significance; these omissions are load-bearing for the central empirical assertion.

Authors: We agree that the abstract would be strengthened by greater specificity on the empirical claims. In the revised manuscript, we will update the abstract to name the primary baseline compilers (Qiskit and t|ket>) and their optimization settings, while noting that full details—including 10 independent runs per circuit, standard deviations, and statistical significance—are reported in Section 4. This change preserves abstract length while making the central results more transparent and verifiable. revision: yes
Referee: [Experiments] Experiments section: no cross-dataset hold-out, no scaling curves with qubit count or depth, and no topology-agnostic test set are described, so it is impossible to determine whether the reported gains arise from the dynamic-QAP formulation or from statistical similarity between the training and test circuit ensembles.

Authors: This is a fair observation on experimental controls. The original evaluation uses three distinct datasets (MQTBench, AgentQ, QUEKO) and multiple hardware topologies, but does not include explicit cross-dataset hold-out, scaling curves, or a dedicated topology-agnostic test set. We will revise the Experiments section to add: (i) scaling curves for qubit count (up to 50) and depth, (ii) a held-out cross-dataset evaluation protocol, and (iii) a topology-agnostic test set. These additions will clarify generalization and isolate the contribution of the dynamic QAP formulation. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical RL method validated externally

full rationale

The paper formulates qubit routing as a dynamic QAP with flow/distance matrices defining the RL reward, then trains a solution-aware Transformer policy with lookahead. All reported CNOT reductions (15.7-30.4%) are measured against independent industry compilers on external benchmark suites (MQTBench, AgentQ, QUEKO), not derived from or forced by any internal parameter fit, self-citation, or definitional equivalence. The QAP objective is an input modeling choice whose outputs are tested rather than presupposed.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard domain assumptions in quantum compilation and reinforcement learning; no new physical entities are postulated and no parameters are fitted to the final CNOT metric itself.

free parameters (1)

RL training hyperparameters
Learning rate, discount factor, and network sizes are required for any RL method but are not reported in the abstract.

axioms (2)

domain assumption Qubit routing is NP-hard and local decisions compound globally
Invoked in the first paragraph to motivate the need for a structured global objective.
domain assumption The QAP objective (flow-distance product) is an appropriate proxy for total swap cost
Used to define the RL reward without further justification in the abstract.

pith-pipeline@v0.9.0 · 5538 in / 1235 out tokens · 51987 ms · 2026-05-13T04:41:13.759262+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

78 extracted references · 78 canonical work pages · 6 internal anchors

[1]

Challenges and opportunities in quantum optimization.Nature Reviews Physics, pages 1–18, 2024

Amira Abbas, Andris Ambainis, Brandon Augustino, Andreas Bärtschi, Harry Buhrman, Carleton Coffrin, Giorgio Cortiana, Vedran Dunjko, Daniel J Egger, Bruce G Elmegreen, et al. Challenges and opportunities in quantum optimization.Nature Reviews Physics, pages 1–18, 2024

work page 2024
[2]

Deep Learning using Rectified Linear Units (ReLU)

Abien Fred Agarap. Deep learning using rectified linear units (relu).arXiv preprint arXiv:1803.08375, 2018

work page internal anchor Pith review Pith/arXiv arXiv 2018
[3]

Layer Normalization

Jimmy Lei Ba, Jamie Ryan Kiros, and Geoffrey E Hinton. Layer normalization.arXiv preprint arXiv:1607.06450, 2016

work page internal anchor Pith review Pith/arXiv arXiv 2016
[4]

Efficient compilation for shuttling trapped-ion machines via the position graph architectural abstraction.arXiv preprint arXiv:2501.12470, 2025

Bao Bach, Ilya Safro, and Ed Younis. Efficient compilation for shuttling trapped-ion machines via the position graph architectural abstraction.arXiv preprint arXiv:2501.12470, 2025

work page arXiv 2025
[5]

Solving the quadratic assignment problem using deep reinforcement learning.arXiv preprint arXiv:2310.01604, 2023

Puneet S Bagga and Arthur Delarue. Solving the quadratic assignment problem using deep reinforcement learning.arXiv preprint arXiv:2310.01604, 2023

work page arXiv 2023
[6]

Parameterized quantum circuits as machine learning models.Quantum Science and Technology, 4(4):043001, 2019

Marcello Benedetti, Erika Lloyd, Stefan Sack, and Mattia Fiorentini. Parameterized quantum circuits as machine learning models.Quantum Science and Technology, 4(4):043001, 2019

work page 2019
[7]

Machine learning for combinatorial optimization: a methodological tour d’horizon, 2020

Yoshua Bengio, Andrea Lodi, and Antoine Prouvost. Machine learning for combinatorial optimization: a methodological tour d’horizon, 2020

work page 2020
[8]

Quantum machine learning.Nature, 549(7671):195, 2017

Jacob Biamonte, Peter Wittek, Nicola Pancotti, Patrick Rebentrost, Nathan Wiebe, and Seth Lloyd. Quantum machine learning.Nature, 549(7671):195, 2017

work page 2017
[9]

Quantum amplitude amplification and estimation

Gilles Brassard, Peter Høyer, Michele Mosca, and Alain Tapp. Quantum amplitude amplification and estimation. InQuantum Computation and Information, volume 305 ofContemporary Mathematics, pages 53–74. American Mathematical Society, 2002

work page 2002
[10]

Openai gym, 2016

Greg Brockman, Vicki Cheung, Ludwig Pettersson, Jonas Schneider, John Schulman, Jie Tang, and Wojciech Zaremba. Openai gym, 2016

work page 2016
[11]

Challenges and opportunities in quantum machine learning.Nature computational science, 2(9):567–576, 2022

Marco Cerezo, Guillaume Verdon, Hsin-Yuan Huang, Lukasz Cincio, and Patrick J Coles. Challenges and opportunities in quantum machine learning.Nature computational science, 2(9):567–576, 2022

work page 2022
[12]

Quantum algorithms revisited.Proceedings of the Royal Society of London

Richard Cleve, Artur Ekert, Chiara Macchiavello, and Michele Mosca. Quantum algorithms revisited.Proceedings of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences, 454(1969):339–354, 1998

work page 1969
[13]

On the qubit routing problem.arXiv preprint arXiv:1902.08091, 2019

Alexander Cowtan, Silas Dilkes, Ross Duncan, Alexandre Krajenbrink, Will Simmons, and Seyon Sivarajah. On the qubit routing problem.arXiv preprint arXiv:1902.08091, 2019

work page arXiv 1902
[14]

Rapid solution of problems by quantum computation

David Deutsch and Richard Jozsa. Rapid solution of problems by quantum computation. Proceedings of the Royal Society of London. Series A: Mathematical and Physical Sciences, 439(1907):553–558, 1992

work page 1907
[15]

Ignacio Cirac

Wolfgang Dür, Guifré Vidal, and J. Ignacio Cirac. Three qubits can be entangled in two inequivalent ways.Physical Review A, 62(6):062314, 2000

work page 2000
[16]

A Quantum Approximate Optimization Algorithm

Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. A quantum approximate optimization algorithm.arXiv preprint arXiv:1411.4028, 2014

work page internal anchor Pith review Pith/arXiv arXiv 2014
[17]

A quantum approximate optimization algorithm

Edward Farhi, Jeffrey Goldstone, and Sam Gutmann. A quantum approximate optimization algorithm. 2014

work page 2014
[18]

Greenberger, Michael A

Daniel M. Greenberger, Michael A. Horne, and Anton Zeilinger. Going beyond Bell’s theorem. In Menas Kafatos, editor,Bell’s Theorem, Quantum Theory and Conceptions of the Universe, pages 69–72. Kluwer Academic Publishers, Dordrecht, 1989. 10

work page 1989
[19]

Enrico Fermi

Marc Hein, Wolfgang Dür, Jens Eisert, Robert Raussendorf, Maarten Van den Nest, and Hans J. Briegel. Entanglement in graph states and its applications. In Giulio Casati, Dima L. Shepelyansky, Peter Zoller, and Giuliano Benenti, editors,Quantum Computers, Algorithms and Chaos, volume 162 ofProceedings of the International School of Physics “Enrico Fermi”, ...

work page 2006
[20]

Using reinforcement learning to find efficient qubit routing policies for deployment in near-term quantum computers.arXiv preprint arXiv:1812.11619, 2018

Steven Herbert and Akash Sengupta. Using reinforcement learning to find efficient qubit routing policies for deployment in near-term quantum computers.arXiv preprint arXiv:1812.11619, 2018

work page arXiv 2018
[21]

Quantum computing for finance.Nature Reviews Physics, 5(8):450–465, 2023

Dylan Herman, Cody Googin, Xiaoyuan Liu, Yue Sun, Alexey Galda, Ilya Safro, Marco Pistoia, and Yuri Alexeev. Quantum computing for finance.Nature Reviews Physics, 5(8):450–465, 2023

work page 2023
[22]

Generic topology mapping strategies for large-scale parallel architectures

Torsten Hoefler and Marc Snir. Generic topology mapping strategies for large-scale parallel architectures. InProceedings of the international conference on Supercomputing, pages 75–84, 2011

work page 2011
[23]

Reinforcement learning and dear framework for solving the qubit mapping problem

Ching-Yao Huang, Chi-Hsiang Lien, and Wai-Kei Mak. Reinforcement learning and dear framework for solving the qubit mapping problem. InProceedings of the 41st IEEE/ACM international conference on computer-aided design, pages 1–9, 2022

work page 2022
[24]

BasicSwap: Qiskit transpiler pass documentation

IBM Quantum. BasicSwap: Qiskit transpiler pass documentation. https://quantum.cloud. ibm.com/docs/api/qiskit/qiskit.transpiler.passes.BasicSwap, 2026. Accessed: 2026-05-05

work page 2026
[25]

EfficientSU2: Qiskit circuit library documentation

IBM Quantum. EfficientSU2: Qiskit circuit library documentation. https://quantum.cloud. ibm.com/docs/api/qiskit/qiskit.circuit.library.EfficientSU2, 2026. Ac- cessed: 2026-05-05

work page 2026
[26]

RealAmplitudes: Qiskit circuit library documentation

IBM Quantum. RealAmplitudes: Qiskit circuit library documentation. https://quantum. cloud.ibm.com/docs/api/qiskit/qiskit.circuit.library.RealAmplitudes,

work page
[27]

Accessed: 2026-05-05

work page 2026
[28]

TwoLocal: Qiskit circuit library documentation

IBM Quantum. TwoLocal: Qiskit circuit library documentation. https://quantum. cloud.ibm.com/docs/api/qiskit/qiskit.circuit.library.TwoLocal, 2026. Ac- cessed: 2026-05-05

work page 2026
[29]

Algorithmic theory of qubit routing

Takehiro Ito, Naonori Kakimura, Naoyuki Kamiyama, Yusuke Kobayashi, and Yoshio Okamoto. Algorithmic theory of qubit routing. InAlgorithms and Data Structures Symposium, pages 533–546. Springer, 2023

work page 2023
[30]

Quantum computing with Qiskit

Ali Javadi-Abhari, Matthew Treinish, Kevin Krsulich, Christopher J Wood, Jake Lishman, Julien Gacon, Simon Martiel, Paul D Nation, Lev S Bishop, Andrew W Cross, et al. Quantum computing with qiskit.arXiv preprint arXiv:2405.08810, 2024

work page internal anchor Pith review Pith/arXiv arXiv 2024
[31]

Agent-q: fine-tuning large language models for quantum circuit generation and optimization

Linus Jern, Valter Uotila, Cong Yu, and Bo Zhao. Agent-q: fine-tuning large language models for quantum circuit generation and optimization. In2025 IEEE International Conference on Quantum Computing and Engineering (QCE), volume 1, pages 1621–1632. IEEE, 2025

work page 2025
[32]

A comprehensive review of quantum circuit optimization: Current trends and future directions

Krishnageetha Karuppasamy, Varun Puram, Stevens Johnson, and Johnson P Thomas. A comprehensive review of quantum circuit optimization: Current trends and future directions. Quantum Reports, 7(1):2, 2025

work page 2025
[33]

Evidence for the utility of quantum computing before fault tolerance.Nature, 618(7965):500–505, 2023

Youngseok Kim, Andrew Eddins, Sajant Anand, Ken Xuan Wei, Ewout Van Den Berg, Sami Rosenblatt, Hasan Nayfeh, Yantao Wu, Michael Zaletel, Kristan Temme, et al. Evidence for the utility of quantum computing before fault tolerance.Nature, 618(7965):500–505, 2023

work page 2023
[34]

Practical and efficient quantum circuit synthesis and transpiling with reinforcement learning

David Kremer, Victor Villar, Hanhee Paik, Ivan Duran, Ismael Faro, and Juan Cruz-Benito. Practical and efficient quantum circuit synthesis and transpiling with reinforcement learning. arXiv preprint arXiv:2405.13196, 2024

work page arXiv 2024
[35]

The quadratic assignment problem.Management science, 9(4):586–599, 1963

Eugene L Lawler. The quadratic assignment problem.Management science, 9(4):586–599, 1963. 11

work page 1963
[36]

The generalized quadratic assignment problem.Research Rep., Dept., Mechanical Industrial Eng., Univ

Chi-Guhn Lee and Zhong Ma. The generalized quadratic assignment problem.Research Rep., Dept., Mechanical Industrial Eng., Univ. Toronto, Canada, page M5S, 2004

work page 2004
[37]

Tackling the qubit mapping problem for nisq-era quantum devices

Gushu Li, Yufei Ding, and Yuan Xie. Tackling the qubit mapping problem for nisq-era quantum devices. InProceedings of the twenty-fourth international conference on architectural support for programming languages and operating systems, pages 1001–1014, 2019

work page 2019
[38]

Post-quantum security: Opportunities and challenges.Sensors, 23(21):8744, 2023

Silong Li, Yuxiang Chen, Lin Chen, Jing Liao, Chanchan Kuang, Kuanching Li, Wei Liang, and Naixue Xiong. Post-quantum security: Opportunities and challenges.Sensors, 23(21):8744, 2023

work page 2023
[39]

Fujitsu and riken develop world-leading 256-qubit superconducting quan- tum computer, 2025

Fujitsu Limited. Fujitsu and riken develop world-leading 256-qubit superconducting quan- tum computer, 2025. Date accessed: May 03, 2026. https://info.archives.global. fujitsu/global/about/resources/news/press-releases/2025/0422-01.html

work page 2025
[40]

Revocable deep reinforcement learning with affinity regularization for outlier-robust graph matching.arXiv preprint arXiv:2012.08950, 2020

Chang Liu, Zetian Jiang, Runzhong Wang, Junchi Yan, Lingxiao Huang, and Pinyan Lu. Revocable deep reinforcement learning with affinity regularization for outlier-robust graph matching.arXiv preprint arXiv:2012.08950, 2020

work page arXiv 2012
[41]

Leveraging special-purpose hardware for local search heuristics.Computational Optimization and Applications, 82(1):1–29, 2022

Xiaoyuan Liu, Hayato Ushijima-Mwesigwa, Avradip Mandal, Sarvagya Upadhyay, Ilya Safro, and Arnab Roy. Leveraging special-purpose hardware for local search heuristics.Computational Optimization and Applications, 82(1):1–29, 2022

work page 2022
[42]

Ising formulations of many np problems.Frontiers in physics, 2:74887, 2014

Andrew Lucas. Ising formulations of many np problems.Frontiers in physics, 2:74887, 2014

work page 2014
[43]

Quantum compiling

Marco Maronese, Lorenzo Moro, Lorenzo Rocutto, and Enrico Prati. Quantum compiling. In Quantum Computing Environments, pages 39–74. Springer, 2022

work page 2022
[44]

Mitarai, M

K. Mitarai, M. Negoro, M. Kitagawa, and K. Fujii. Quantum circuit learning.Phys. Rev. A, 98(3):032309, 2018

work page 2018
[45]

Generating compilers for qubit mapping and routing.Proceedings of the ACM on Programming Languages, 10(POPL):2265–2294, 2026

Abtin Molavi, Amanda Xu, Ethan Cecchetti, Swamit Tannu, and Aws Albarghouthi. Generating compilers for qubit mapping and routing.Proceedings of the ACM on Programming Languages, 10(POPL):2265–2294, 2026

work page 2026
[46]

Qubit mapping and routing via maxsat

Abtin Molavi, Amanda Xu, Martin Diges, Lauren Pick, Swamit Tannu, and Aws Albarghouthi. Qubit mapping and routing via maxsat. In2022 55th IEEE/ACM international symposium on Microarchitecture (MICRO), pages 1078–1091. IEEE, 2022

work page 2022
[47]

Noise-adaptive compiler mappings for noisy intermediate-scale quantum com- puters

Prakash Murali, Jonathan M Baker, Ali Javadi-Abhari, Frederic T Chong, and Margaret Martonosi. Noise-adaptive compiler mappings for noisy intermediate-scale quantum com- puters. InProceedings of the twenty-fourth international conference on architectural support for programming languages and operating systems, pages 1015–1029, 2019

work page 2019
[48]

A hardware-aware heuristic for the qubit mapping problem in the nisq era.IEEE Transactions on Quantum Engineering, 1:1–14, 2020

Siyuan Niu, Adrien Suau, Gabriel Staffelbach, and Aida Todri-Sanial. A hardware-aware heuristic for the qubit mapping problem in the nisq era.IEEE Transactions on Quantum Engineering, 1:1–14, 2020

work page 2020
[49]

Revised note on learning quadratic assignment with graph neural networks

Alex Nowak, Soledad Villar, Afonso S Bandeira, and Joan Bruna. Revised note on learning quadratic assignment with graph neural networks. In2018 IEEE Data Science Workshop (DSW), pages 1–5. IEEE, 2018

work page 2018
[50]

Pytorch: An imperative style, high-performance deep learning library.Advances in neural information processing systems, 32, 2019

Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, et al. Pytorch: An imperative style, high-performance deep learning library.Advances in neural information processing systems, 32, 2019

work page 2019
[51]

A variational eigenvalue solver on a photonic quantum processor.Nature communications, 5:4213, 2014

Alberto Peruzzo, Jarrod McClean, Peter Shadbolt, Man-Hong Yung, Xiao-Qi Zhou, Peter J Love, Alán Aspuru-Guzik, and Jeremy L O’brien. A variational eigenvalue solver on a photonic quantum processor.Nature communications, 5:4213, 2014

work page 2014
[52]

Security in quantum cryptography.Reviews of Modern Physics, 94(2):025008, 2022

Christopher Portmann and Renato Renner. Security in quantum cryptography.Reviews of Modern Physics, 94(2):025008, 2022. 12

work page 2022
[53]

Using reinforcement learning to perform qubit routing in quantum compilers.ACM Transactions on Quantum Computing, 3(2):1–25, 2022

Matteo G Pozzi, Steven J Herbert, Akash Sengupta, and Robert D Mullins. Using reinforcement learning to perform qubit routing in quantum compilers.ACM Transactions on Quantum Computing, 3(2):1–25, 2022

work page 2022
[54]

MQT Bench: Benchmarking software and design automation tools for quantum computing.Quantum, 2023

Nils Quetschlich, Lukas Burgholzer, and Robert Wille. MQT Bench: Benchmarking software and design automation tools for quantum computing.Quantum, 2023. MQT Bench is available athttps://www.cda.cit.tum.de/mqtbench/

work page 2023
[55]

Stable-baselines3: Reliable reinforcement learning implementations.Journal of machine learning research, 22(268):1–8, 2021

Antonin Raffin, Ashley Hill, Adam Gleave, Anssi Kanervisto, Maximilian Ernestus, and Noah Dormann. Stable-baselines3: Reliable reinforcement learning implementations.Journal of machine learning research, 22(268):1–8, 2021

work page 2021
[56]

Robert Raussendorf and Hans J. Briegel. A one-way quantum computer.Physical Review Letters, 86(22):5188–5191, 2001

work page 2001
[57]

Proximal Policy Optimization Algorithms

John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms.arXiv preprint arXiv:1707.06347, 2017

work page internal anchor Pith review Pith/arXiv arXiv 2017
[58]

Better process mapping and sparse quadradic assignment, 2017

Christian Schulz and Jesper Larsson Träff. Better process mapping and sparse quadradic assignment, 2017

work page 2017
[59]

Network community detection on small quantum computers.Advanced Quantum Technologies, 2(9):1900029, 2019

Ruslan Shaydulin, Hayato Ushijima-Mwesigwa, Ilya Safro, Susan Mniszewski, and Yuri Alexeev. Network community detection on small quantum computers.Advanced Quantum Technologies, 2(9):1900029, 2019

work page 2019
[60]

Qubit routing using graph neural network aided monte carlo tree search

Animesh Sinha, Utkarsh Azad, and Harjinder Singh. Qubit routing using graph neural network aided monte carlo tree search. InProceedings of the AAAI conference on artificial intelligence, volume 36, pages 9935–9943, 2022

work page 2022
[61]

Qubit allocation

Marcos Siraichi, Vinicius Fernandes Dos Santos, Sylv ain Collange, and Fernando Magno Quin- tão Pereira. Qubit allocation. InCGO 2018-IEEE/ACM International Symposium on Code Generation and Op timization, pages 1–12, 2018

work page 2018
[62]

t| ket>: a retargetable compiler for nisq devices.Quantum Science & Technology, 6(1):014003, 2021

Seyon Sivarajah, Silas Dilkes, Alexander Cowtan, Will Simmons, Alec Edgington, and Ross Duncan. t| ket>: a retargetable compiler for nisq devices.Quantum Science & Technology, 6(1):014003, 2021

work page 2021
[63]

An open-source, industrial-strength optimizing compiler for quantum programs.Quantum Science & Technology, 5(4):044001, 2020

Robert S Smith, Eric C Peterson, Mark G Skilbeck, and Erik J Davis. An open-source, industrial-strength optimizing compiler for quantum programs.Quantum Science & Technology, 5(4):044001, 2020

work page 2020
[64]

Optimization of the hop-byte metric for effective topol- ogy aware mapping

C Devi Sudheer and Ashok Srinivasan. Optimization of the hop-byte metric for effective topol- ogy aware mapping. In2012 19th International Conference on High Performance Computing, pages 1–9. IEEE, 2012

work page 2012
[65]

Optimality study of existing quantum computing layout synthesis tools.IEEE Transactions on Computers, 70(9):1363–1373, 2020

Bochen Tan and Jason Cong. Optimality study of existing quantum computing layout synthesis tools.IEEE Transactions on Computers, 70(9):1363–1373, 2020

work page 2020
[66]

Learning solution-aware transformers for efficiently solving quadratic assignment problem.arXiv preprint arXiv:2406.09899, 2024

Zhentao Tan and Yadong Mu. Learning solution-aware transformers for efficiently solving quadratic assignment problem.arXiv preprint arXiv:2406.09899, 2024

work page arXiv 2024
[67]

Alpharouter: Quantum circuit routing with reinforcement learning and tree search

Wei Tang, Yiheng Duan, Yaroslav Kharkov, Rasool Fakoor, Eric Kessler, and Yunong Shi. Alpharouter: Quantum circuit routing with reinforcement learning and tree search. In2024 IEEE International Conference on Quantum Computing and Engineering (QCE), volume 1, pages 930–940. IEEE, 2024

work page 2024
[68]

An ising-based model for qubit mapping

Hayato Ushijima Mwesigwa and Xiaoyuan Liu. An ising-based model for qubit mapping. In Proceedings of the SC’23 Workshops of the International Conference on High Performance Computing, Network, Storage, and Analysis, pages 1492–1498, 2023

work page 2023
[69]

Attention is all you need

Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need. InAdvances in neural information processing systems, pages 5998–6008, 2017. 13

work page 2017
[70]

Runzhong Wang, Junchi Yan, and Xiaokang Yang. Neural graph matching network: Learn- ing lawler’s quadratic assignment problem with extension to hypergraph and multiple-graph matching.IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(9):5261–5279, 2021

work page 2021
[71]

Implementation of the quantum fourier transform.Physical review letters, 86(9):1889, 2001

Yaakov S Weinstein, MA Pravia, EM Fortunato, Seth Lloyd, and David G Cory. Implementation of the quantum fourier transform.Physical review letters, 86(9):1889, 2001

work page 2001
[72]

Quantum circuit synthesis and compilation optimization: Overview and prospects.arXiv preprint arXiv:2407.00736, 2024

Ge Yan, Wenjie Wu, Yuheng Chen, Kaisen Pan, Xudong Lu, Zixiang Zhou, Yuhan Wang, Ruocheng Wang, and Junchi Yan. Quantum circuit synthesis and compilation optimization: Overview and prospects.arXiv preprint arXiv:2407.00736, 2024

work page arXiv 2024
[73]

Learning local search heuristics for boolean satisfiability

Emre Yolcu and Barnabás Póczos. Learning local search heuristics for boolean satisfiability. Advances in Neural Information Processing Systems, 32, 2019

work page 2019
[74]

Quantum machine learning: A review and case studies.Entropy, 25(2):287, 2023

Amine Zeguendry, Zahi Jarir, and Mohamed Quafafou. Quantum machine learning: A review and case studies.Entropy, 25(2):287, 2023

work page 2023
[75]

Exponential quantum advantage in processing massive classical data

Haimeng Zhao, Alexander Zlokapa, Hartmut Neven, Ryan Babbush, John Preskill, Jarrod R McClean, and Hsin-Yuan Huang. Exponential quantum advantage in processing massive classical data.arXiv preprint arXiv:2604.07639, 2026

work page internal anchor Pith review Pith/arXiv arXiv 2026
[76]

Lightsabre: A lightweight and enhanced sabre algorithm.arXiv preprint arXiv:2409.08368, 2024

Henry Zou, Matthew Treinish, Kevin Hartman, Alexander Ivrii, and Jake Lishman. Lightsabre: A lightweight and enhanced sabre algorithm.arXiv preprint arXiv:2409.08368, 2024

work page arXiv 2024
[77]

look-ahead

Alwin Zulehner, Alexandru Paler, and Robert Wille. An efficient methodology for mapping quantum circuits to the ibm qx architectures.IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, 2018. A Related Work Qubit mapping and routing are of central importance in quantum compilation. These problems are often defined in terms of boo...

work page 2018
[78]

This bidirectional procedure allows the router to account for both early and late two-qubit interactions

Finally, a third forward pass routes the original circuit again using the layout information obtained from the backward passX ′′ 0. This bidirectional procedure allows the router to account for both early and late two-qubit interactions. Compared with applying a single forward pass, the forward–backward–forward refinement can reduce unnecessary SWAP inser...

work page 2000