Pith. sign in

REVIEW 3 major objections 3 minor 1 cited by

Unitary-control-then-measure quantum RL models cut expected-return complexity from exponential in horizon length to a power law, and their optimal policies show measurement-driven degeneracies absent from ordinary quantum control.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-12 23:48 UTC pith:FC6TQZBJ

load-bearing objection Solvable unitary-control-then-measure QRL models with a claimed exp-to-power-law complexity drop that still leans on a conjecture, plus measurement-induced policy degeneracy that looks real in the four-level case. the 3 major comments →

arxiv 2604.13096 v2 pith:FC6TQZBJ submitted 2026-04-09 math.GM

Complexity scaling and optimal policy degeneracy in quantum reinforcement learning via analytically solvable unitary-control-then-measure models

Andrea Cintio , Alessandro Michelangeli , Dmitrii Tsutskov This is my paper
classification math.GM MSC 81P6868T0590C40
keywords quantum reinforcement learningunitary-control-then-measureMarkov decision processcomplexity scalingpolicy degeneracyquantum Zeno effecttrajectory equivalencefinite-horizon MDP
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper constructs a family of exactly solvable finite-horizon quantum reinforcement-learning models in which an agent applies unitary controls interleaved with projective measurements onto a fixed basis. Closed-form trajectory probabilities, rewards and expected returns are obtained for four concrete systems: two qubit chains, a ladder-coupled qutrit and a four-level two-qubit system. These expressions reveal that the nominal exponential cost of summing over trajectories collapses to a power-law scaling, because many trajectories are equivalent, the transition graph is sparse, and (conjecturally) the optimal return concentrates on a polynomially sized set of classes. At the same time the models exhibit unique Zeno-governed optima in low dimension and both plateau-type and discrete degeneracies of the optimal policy in the four-level system—features with no counterpart in measurement-free quantum optimal control. A sympathetic reader cares because the constructions supply the first analytically transparent laboratory in which complexity scaling and policy degeneracy of quantum RL can be quantified exactly rather than simulated.

Core claim

In the proposed unitary-control-then-measure quantum reinforcement-learning protocols the computational complexity of the expected return is reduced from the nominally exponential O(e^N) scaling in trajectory length N to an explicit power-law O(N^I). The reduction is driven by two rigorously established mechanisms—trajectory equivalence and sparsity of the transition graph—together with a third, conjectured spectral concentration of the return onto polynomially populated trajectory classes at the optimal policy. Concurrently the optimal policies themselves display measurement-induced degeneracies (Zeno asymptotics, plateaus and discrete critical degeneracies) that have no analogue in the mea

What carries the argument

The unitary-control-then-measure protocol: each decision step consists of a unitary transformation of the quantum state followed by a projective measurement onto a prescribed reference basis, turning the learning problem into a finite-horizon Markov decision process whose trajectory probabilities admit exact closed forms.

Load-bearing premise

The third complexity-reduction mechanism—spectral concentration of the return onto polynomially many trajectory classes at optimality—is only conjectured, and the four low-dimensional realisations are taken to be representative of broader quantum reinforcement learning.

What would settle it

Compute the expected-return sum for one of the four models at large N both by enumerating all trajectories and by the claimed power-law reduction; if the numerical cost remains exponential or the optimal return fails to concentrate on the predicted polynomial classes, the central complexity claim fails.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Exact benchmarking of approximate quantum-RL algorithms becomes possible against closed-form returns rather than Monte-Carlo estimates.
  • Power-law rather than exponential scaling of the return suggests that certain quantum RL problems remain tractable at horizons previously regarded as intractable.
  • Measurement-induced policy degeneracies must be accounted for when designing or certifying quantum controllers; uniqueness assumptions inherited from measurement-free optimal control no longer hold.
  • The quantum Zeno effect can be harnessed as a design principle that forces optimal policies toward continuous monitoring regimes as N grows.

Where Pith is reading between the lines

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

  • If the conjectured spectral concentration generalises beyond the four models, quantum RL with intermediate measurements may systematically evade the curse of dimensionality that plagues classical trajectory-based RL.
  • The discrete degeneracies at critical energy parameters hint that phase-transition-like phenomena may organise the policy landscape of quantum RL, inviting a statistical-mechanics treatment of the return functional.
  • Extending the same unitary-then-measure construction to continuous-variable or many-body systems would test whether the power-law reduction survives Hilbert-space dimension growth.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 3 minor

Summary. The manuscript proposes a class of analytically solvable quantum reinforcement learning (QRL) models formulated as finite-horizon MDPs on finite-dimensional Hilbert spaces under a unitary-control-then-measure protocol (unitary controls interleaved with projective measurements onto a fixed reference basis). Exact closed-form expressions for trajectory probabilities, rewards and expected return are claimed for four low-dimensional realisations (closed-chain and anti-periodic qubits, ladder-coupled qutrit, four-level two-qubit system). Two structural results are then asserted: (i) a reduction of expected-return complexity from the nominal O(e^N) trajectory scaling to an explicit power-law O(N^I), attributed to trajectory equivalence and transition-graph sparsity (claimed rigorous) together with a conjectured spectral concentration of the return onto polynomially populated classes at the optimal policy; (ii) characterisation of optimal-policy degeneracy, with unique Zeno-governed optima in the low-dimensional cases and both plateau-type quasi-degeneracy and genuine discrete degeneracy (at critical energies) in the four-level system, phenomena absent from measurement-free quantum optimal control.

Significance. Exact solvability is rare in QRL; if the closed forms and the two rigorous complexity mechanisms hold, the work supplies a useful analytical laboratory for studying return evaluation and policy structure under measurement. The explicit identification of trajectory equivalence and graph sparsity as complexity-reduction mechanisms, and the contrast with measurement-free landscapes for degeneracy, are potentially valuable. The four concrete models and any machine-checkable derivations or reproducible numerics would strengthen the contribution. Broader claims for QRL, however, rest on low-dimensional examples and on a still-conjectural third mechanism, so the significance remains conditional on clarifying the precise scope of the proven power-law reduction.

major comments (3)
  1. [Abstract / complexity-reduction analysis] Abstract and complexity section: the headline claim that expected-return complexity falls from O(e^N) to O(N^I) is said to be driven by two rigorously established mechanisms (trajectory equivalence, transition-graph sparsity) plus a third, explicitly conjectured spectral concentration. The manuscript never states the scaling of the number of equivalence classes after the two rigorous mechanisms alone. If that count remains exponential (or super-polynomial), the power-law evaluation complexity is conditional on the unproven concentration; if the two mechanisms already yield only polynomially many classes, the conjecture is superfluous as a co-driver and should not be listed as such. A precise statement of the class-count scaling under equivalence+sparsity, and a clean separation of what is proven versus conjectural for the O(N^I) claim, is load-bearing.
  2. [Complexity scaling claims] The exponent I and the concrete scaling of the number of trajectory classes are not quantified in the abstract summary of results; without an explicit expression or bound for I (or for the class count after the rigorous reductions), the power-law claim cannot be verified as stated and the computational advantage remains unquantified.
  3. [Optimal-policy degeneracy / four-level system] The structural features (Zeno asymptotics of unique optima; plateau and discrete degeneracy) are established only for four low-dimensional realisations. The manuscript does not supply evidence or argument that these models are representative of higher-dimensional or less structured unitary-control-then-measure protocols, so the claim that the phenomena are characteristic of the QRL protocol class more broadly is under-supported.
minor comments (3)
  1. [Complexity analysis] Clarify the precise definition of the dimension-dependent exponent I and whether it is model-independent or depends on the Hilbert-space dimension / coupling structure of each realisation.
  2. [Front matter] The primary category math.GM is atypical for a QRL/quantum-control paper; ensure the abstract and introduction make the mathematical contribution (exact solvability, complexity reduction) transparent to a general-mathematics audience.
  3. [Results summary] When the full proofs and any numerical checks of the concentration conjecture are present, add a short table or remark that lists, for each of the four models, the proven class-count scaling versus the residual role of the conjecture.

Circularity Check

0 steps flagged

No significant circularity: closed-form analysis of defined unitary-control-then-measure models, not a fit or self-definitional loop.

full rationale

The paper defines a unitary-control-then-measure protocol as finite-horizon MDPs on finite-dimensional Hilbert spaces, then derives exact closed-form trajectory probabilities, rewards, and expected returns for four concrete low-dimensional realisations (closed-chain and anti-periodic qubit, ladder qutrit, four-level two-qubit). From those expressions it analyses two structural features: (i) reduction of expected-return complexity from nominal O(e^N) to power-law O(N^I), attributed to trajectory equivalence and transition-graph sparsity (stated as rigorous) plus an openly conjectured spectral concentration at the optimal policy, and (ii) optimal-policy degeneracy (Zeno asymptotics in low-dim models; plateau and discrete degeneracy in the four-level system). Nothing in the available text equates a claimed prediction to a fitted parameter, defines the result in terms of itself, imports a uniqueness theorem from overlapping authors as an external fact that forces the claim, or renames a known empirical pattern as a new derivation. The third complexity mechanism is explicitly labelled conjectured rather than smuggled in as proven. Residual risk is ordinary mathematical correctness of the closed forms and whether equivalence+sparsity alone already yield polynomially many classes (a support/completeness issue, not circularity). The derivation chain is self-contained theoretical analysis of defined models; score 0 with empty steps is the honest finding.

Axiom & Free-Parameter Ledger

2 free parameters · 4 axioms · 1 invented entities

Abstract-only review: free parameters, axioms, and invented entities are inferred from the stated modeling choices. The load-bearing setup is a finite-horizon MDP on a finite-dimensional Hilbert space with a unitary-control-then-measure protocol and projective measurements onto a fixed basis. Concrete realisations (closed-chain and anti-periodic qubits, ladder qutrit, four-level two-qubit) and energy parameters that trigger discrete degeneracy are model choices. The third complexity mechanism is explicitly conjectural. No empirical fits are claimed.

free parameters (2)
  • Energy parameters of the four-level system (critical values for discrete degeneracy)
    Abstract states genuine discrete degeneracy at critical energy parameters; those critical values are model parameters that select the degeneracy phenomenon.
  • Horizon length N and dimension-dependent exponent I in O(N^I)
    N is the trajectory length; I is the power in the claimed complexity scaling and depends on the model structure. Not fitted to external data but central to the complexity claim.
axioms (4)
  • domain assumption QRL is formulated as a finite-horizon Markov decision process on a finite-dimensional Hilbert space.
    Stated framing of the entire class of models; standard MDP structure plus finite quantum state space.
  • ad hoc to paper Agent protocol is unitary control interleaved with projective measurement onto a prescribed reference basis (unitary-control-then-measure).
    Defines the solvable class; not forced by general QRL, chosen to enable closed forms.
  • ad hoc to paper Trajectory equivalence and transition-graph sparsity rigorously reduce expected-return complexity; spectral concentration at the optimal policy is conjectured.
    Abstract separates two rigorous mechanisms from a third conjectured one; the conjecture is load-bearing for the full power-law story at optimality.
  • standard math Standard quantum measurement and unitary evolution postulates (Born rule, projective collapse, unitary maps).
    Background quantum mechanics needed for trajectory probabilities and Zeno asymptotics.
invented entities (1)
  • Unitary-control-then-measure QRL protocol class (closed-chain qubit, anti-periodic qubit, ladder qutrit, four-level two-qubit realisations) no independent evidence
    purpose: Provide analytically solvable finite-horizon quantum MDPs with closed-form trajectory probabilities, rewards, and expected return.
    These concrete models are introduced as the paper's solvable realisations; independent evidence outside the paper is not established from the abstract alone.

pith-pipeline@v1.1.0-grok45 · 6649 in / 2897 out tokens · 28098 ms · 2026-07-12T23:48:59.111727+00:00 · methodology

0 comments
read the original abstract

We propose and analyse a class of analytically solvable models of quantum reinforcement learning (QRL), formulated as finite-horizon Markov decision processes in finite-dimensional Hilbert spaces. The models are built around a `unitary-control-then-measure' protocol, in which a learning agent applies unitary transformations to a quantum state and interleaves each control step with a projective measurement onto a prescribed reference basis. Exact closed-form expressions for trajectory probabilities, rewards, and the expected return are derived for four concrete realisations: a closed-chain and an anti-periodic qubit implementation, a qutrit model with ladder coupling, and a four-level two-qubit system. Two structural features of these QRL protocols are then analysed. First, we identify and quantify the reduction in the computational complexity of the expected return, from the nominally exponential $O(e^N)$ scaling in the trajectory length~$N$ to an explicit power-law $O(N^{\mathcal{I}})$, driven by two rigorously established mechanisms, a trajectory equivalence and a sparsity of the transition graph, besides a third, conjectured one: a spectral concentration of the return, at the optimal policy, onto the polynomially populated trajectory classes. Second, we characterise the degeneracy of optimal policies. The low-dimensional models exhibit unique optima whose asymptotic behaviour with~$N$ is governed by the quantum Zeno effect, while the four-level system displays both plateau-type quasi-degeneracy at large horizons and genuine discrete degeneracy at critical energy parameters -- phenomena with no counterpart in the measurement-free quantum optimal control landscape.

Figures

Figures reproduced from arXiv: 2604.13096 by Alessandro Michelangeli, Andrea Cintio, Dmitrii Tsutskov.

Figure 1
Figure 1. Figure 1: Numerical optimisation of the Expected Return (4.19) the additional positive return from these terms (specifically from the (p−1)x+ parts), the optimization pushes x− to be strictly positive. • The penalty limits x−: the reward kernel inside the sum contains the negative term −(N + 1 − p)x−. Since this coefficient scales as O(N), x− must scale as O(1/N) to keep the penalty finite (order O(1)). Consequently… view at source ↗
Figure 2
Figure 2. Figure 2: Numerical optimisation of the Expected Return (5.7) Therefore, the expected return for this model is J(π) = X N p=1 min{p X ,N+1−p} c=1  p − 1 c − 1  ·  N − p c − 1  [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Left: the directed graph G3 on node set {|0⟩, |1⟩, |2⟩} with all six off-diagonal edges, whose incidence matrix has rank 3 − 1 = 2 [4, Theo￾rem 2.3]. Right: the reduced transition graph Ge3, obtained from G3 by re￾taining only the three edges |i⟩ → |j⟩ with nonzero probability after imposing the forbidden-transition constraints (6.14); edge labels indicate the transi￾tion multiplicities aij from (6.16), fo… view at source ↗
Figure 4
Figure 4. Figure 4: Numerical optimisation of the Expected Return (6.31) The latter formula has a symmetry that is explicitly seen by restoring n1 from n0+n1+n2 = N + 1. One then re-writes Jε(x) = N X−1 n0=1 N X−n0 n2=1 min{n0X,n1,n2}−1 c=0  n0 − 1 c n1 − 1 c n2 − 1 c  ×  ε(n1 − n0) + (n0 + 2n2 − N − 2) x + 1 (1 − x) N−3c−2x 3c+2 , with n0 + n1 + n2 = N + 1 . (6.30) Now, swapping n0 ↔ n1 leaves the product of the bi… view at source ↗
Figure 5
Figure 5. Figure 5: Optimal policy (x ∗ , y∗ , z∗ ) for the expected return (6.34) as a func￾tion of the horizon N, for ε = 0.75 (left) and ε = 3.0 (right). The saturations y ∗ → 1 and z ∗ → 0 are interpreted in the text. dominates over reward: the gain from eliminating the feedback arc outweighs the foregone (n2 − 1)z term, since each feedback cycle would re-expose the system to the now-dominant ε-penalty at |0⟩. 7. A compar… view at source ↗
Figure 6
Figure 6. Figure 6: Comparison of computational time required to evaluate the ex￾pected return Jε(x) as a function of trajectory length N. The ‘Brute Force’ approach (red dashed line) scales exponentially as O(3N ), while the analytic formula (blue solid line) scales polynomially as O(N3 ). The sum 7.1 runs over the 3 N−1 possible configurations of the intermediate states, and is parametrised by the intermediate energy level … view at source ↗
Figure 7
Figure 7. Figure 7: Left: the directed graph G4 on node set {|0⟩, |1⟩, |1 ′ ⟩, |2⟩} with all twelve off-diagonal edges (i ̸= j), whose incidence matrix has rank 4 − 1 = 3 [4, Theorem 2.3]. Right: the reduced transition graph Ge4, obtained from G4 by retaining only the five edges |i⟩ → |j⟩ with nonzero probability after imposing the forbidden-transition constraints (8.24); edge labels indicate the transition multiplicities aij… view at source ↗
Figure 8
Figure 8. Figure 8: Numerical optimisation of the expected return (8.33) on a horizon of N = 8 long trajectories. The plots show the migration of the optimal policy inside the control space [0, 1]×[0, 1] as the energy penalties (ε, ε′ ) are varied. Remarkably, the linear dependence on the branching fluxes c01 and c01′ cancels out exactly (e.g., the energy cost of entering, traversing, and exiting the upper branch is balanced … view at source ↗
Figure 9
Figure 9. Figure 9: Numerical evidence of the emergence of a plateau-like region in the profile of the return function (8.33) around the point of optimal policy, as N increases [PITH_FULL_IMAGE:figures/full_fig_p029_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Emergence of degenerate optimal policies for the expected re￾turn (8.33) (horizon N = 8). A crossover of the global maximum (xmax, x′ max) between two distinct regions of the domain is observed when increasing ε from 2.47 to 2.48. Since the return Jε,1 is continuous in ε, this implies the existence of a critical value ε ∗ ∈ (2.47, 2.48) where two distinct optimal poli￾cies coexist with J max ε ∗,1 ≈ 0.644… view at source ↗

Review history (2 revisions) →

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. QANTIS: Hardware-Calibrated Sequential POMDP Belief Updates on IBM Heron

    cs.AI 2026-07 conditional novelty 5.0

    All-step fixed-point amplitude amplification on IBM Heron preserves sequential Tiger POMDP posteriors and planner actions across 8–32 step horizons inside a measured operating envelope.

Reference graph

Works this paper leans on

29 extracted references · 1 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Aharonov, L

    Y. Aharonov, L. Davidovich, and N. Zagury , Quantum random walks , Phys. Rev. A, 48 (1993), pp. 1687--1690

    1993

  2. [2]

    Arulkumaran, M

    K. Arulkumaran, M. P. Deisenroth, M. Brundage, and A. A. Bharath , Deep reinforcement learning: A brief survey , IEEE signal processing magazine, 34 (2017), pp. 26--38

    2017

  3. [3]

    Attal, F

    S. Attal, F. Petruccione, C. Sabot, and I. Sinayskiy , Open Quantum Random Walks , J. Stat. Phys., 147 (2012), pp. 832--852

    2012

  4. [4]

    R. B. Bapat , Graphs and Matrices , Universitext , Springer, London, 2 ed., 2014

    2014

  5. [5]

    A. Barr, W. Gispen, and A. Lamacraft , Quantum Ground States from Reinforcement Learning , in Proceedings of Machine Learning Research , vol. 107, PMLR, 2020, pp. 635--653

    2020

  6. [6]

    Bertsekas , Reinforcement learning and optimal control , vol

    D. Bertsekas , Reinforcement learning and optimal control , vol. 1, Athena Scientific, 2019

    2019

  7. [7]

    height 2pt depth -1.6pt width 23pt, A course in reinforcement learning , Athena Scientific, 2024

    2024

  8. [8]

    H. J. Briegel and G. De las Cuevas , Projective simulation for artificial intelligence , Sci. Rep., 2 (2012), p. 400

    2012

  9. [9]

    Clarke and F

    J. Clarke and F. K. Wilhelm , Superconducting quantum bits , Nature, 453 (2008), pp. pages 1031--1042

    2008

  10. [10]

    Cohen-Tannoudji, B

    C. Cohen-Tannoudji, B. Diu, and F. Lalo \"e , Quantum mechanics , Wiley, New York, NY, 1977. Trans. of : M \'e canique quantique. Paris : Hermann, 1973

    1977

  11. [11]

    D. Dong, C. Chen, H. Li, and T.-J. Tarn , Quantum Reinforcement Learning , IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 38 (2008), pp. 1207--1220

    2008

  12. [12]

    1207--1220

    height 2pt depth -1.6pt width 23pt, Quantum reinforcement learning , IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 38 (2008), pp. 1207--1220

    2008

  13. [13]

    Facchi and S

    P. Facchi and S. Pascazio , Quantum Zeno dynamics: mathematical and physical aspects , Journal of Physics A: Mathematical and Theoretical, 41 (2008), p. 493001

    2008

  14. [14]

    Hsieh and H

    M. Hsieh and H. Rabitz , Optimal control landscape for the generation of unitary transformations , Phys. Rev. A, 77 (2008), p. 042306

    2008

  15. [15]

    Hsieh, R

    M. Hsieh, R. Wu, H. Rabitz, and D. Lidar , Optimal control landscape for the generation of unitary transformations with constrained dynamics , Phys. Rev. A, 81 (2010), p. 062352

    2010

  16. [16]

    L. P. Kaelbling, M. L. Littman, and A. W. Moore , Reinforcement learning: A survey , Journal of artificial intelligence research, 4 (1996), pp. 237--285

    1996

  17. [17]

    Leibfried, R

    D. Leibfried, R. Blatt, C. Monroe, and D. Wineland , Quantum dynamics of single trapped ions , Rev. Mod. Phys., 75 (2003), pp. 281--324

    2003

  18. [18]

    Meyer, C

    N. Meyer, C. Ufrecht, M. Periyasamy, D. D. Scherer, A. Plinge, and C. Mutschler , A Survey on Quantum Reinforcement Learning , arXiv:2211.03464 (2022)

    Pith/arXiv arXiv 2022

  19. [19]

    G. D. Paparo, V. Dunjko, A. Makmal, M. A. Martin-Delgado, and H. J. Briegel , Quantum speed-up for active learning agents , Phys. Rev. X, 4 (2014), p. 031002

    2014

  20. [20]

    M. L. Puterman , Markov decision processes: discrete stochastic dynamic programming , John Wiley & Sons, 2014

    2014

  21. [21]

    S. M. Reimann and M. Manninen , Electronic structure of quantum dots , Rev. Mod. Phys., 74 (2002), pp. 1283--1342

    2002

  22. [22]

    Schenk, E

    M. Schenk, E. F. Combarro, M. Grossi, V. Kain, K. S. B. Li, M.-M. Popa, and S. Vallecorsa , Hybrid actor-critic algorithm for quantum reinforcement learning at cern beam lines , Quantum Science and Technology, 9 (2024), p. 025012

    2024

  23. [23]

    Sugny and C

    D. Sugny and C. Kontz , Optimal control of a three-level quantum system by laser fields plus von Neumann measurements , Phys. Rev. A, 77 (2008), p. 063420

    2008

  24. [24]

    R. S. Sutton and A. G. Barto , Reinforcement Learning: An Introduction , The MIT Press, second ed., 2018

    2018

  25. [25]

    Szepesv \'a ri , Algorithms for reinforcement learning , Springer nature, 2022

    C. Szepesv \'a ri , Algorithms for reinforcement learning , Springer nature, 2022

    2022

  26. [26]

    S. E. Venegas-Andraca , Quantum walks: a comprehensive review , Quantum Inf. Process., 11 (2012), pp. 1015--1106

    2012

  27. [27]

    Volkov, A

    B. Volkov, A. Myachkova, and A. Pechen , Phenomenon of a stronger trapping behavior in -type quantum systems with symmetry , Phys. Rev. A, 111 (2025), p. 022617

    2025

  28. [28]

    Wendin , Quantum information processing with superconducting circuits: a review , Reports on Progress in Physics, 80 (2017), p

    G. Wendin , Quantum information processing with superconducting circuits: a review , Reports on Progress in Physics, 80 (2017), p. 106001

    2017

  29. [29]

    S. Wu, S. Jin, D. Wen, D. Han, and X. Wang , Quantum reinforcement learning in continuous action space , Quantum, 9 (2025), p. 1660

    2025