Pith. sign in

REVIEW 5 minor 1 cited by

Randomized Quantum Optimal Control

T0 review · 0 major / 5 minor · reviewed 2026-07-14 · grok-4.5

Pith's one-line read Randomized ensembles of control waveforms can reach a target quantum operation faster, and with quadratically smaller error, than any single deterministic control under the same bounds.

desk verdict Clean existence proof that randomized continuous-time QOC can beat the deterministic optimum under identical constraints, plus usable symmetry constructions that deliver the predicted quadratic error drop on CNOT. read the letter →

arxiv 2607.10946 v1 pith:JHUEFJHF submitted 2026-07-12 quant-ph

classification quant-ph
keywords quantumoptimalcontrolrandomizedcoherent-errorcancellationdiamonddistanceGRAPECNOTboundarypulsesnoiserobustness
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Quantum optimal control usually hunts for one best control waveform that steers a quantum system to a desired gate. This paper instead optimizes over an ensemble of waveforms together with the probabilities of using each one. Because different branches can carry opposite coherent errors, averaging them can cancel the leading error term, leaving only a quadratic remainder. An exactly solvable single-qubit model proves that the best randomized error is exactly the square of the best deterministic error, so a prescribed accuracy is reached in strictly less time. The same cancellation idea yields practical constructions that turn any good deterministic pulse into a randomized ensemble, and numerical CNOT benchmarks confirm the predicted quadratic improvement. The framework also supplies randomized boundary-pulse schemes that cancel first-order coherent noise for Clifford gates.

What carries the argument

Phase-aligned averaged unitary: the ensemble-averaged operator Up,φ whose first-order error generator is the probability-weighted average of the individual branch error generators. When that average vanishes, the diamond-distance error drops from linear to quadratic order.

What would settle it

Outside the short-time regime of the single-qubit model, compute both the deterministic and randomized diamond-distance errors versus total time; if the randomized error fails to stay below the square of the deterministic error (or if the time to reach a fixed intermediate accuracy is no longer shorter), the claimed quadratic separation is falsified.

Watch

Extended reading notes

Core claim

Under identical control constraints there exist systems, targets and accuracy thresholds for which the minimal time needed by a randomized control ensemble is strictly smaller than the minimal time needed by any deterministic waveform; in the solvable single-qubit case the optimal randomized diamond-distance error is exactly the square of the optimal deterministic error.

Load-bearing premise

The rigorous speed-up proof needs a short-time, moderate-amplitude regime in which the target is unreachable and a sign-flipped partner of the unique deterministic optimum exactly reverses its error generator.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

0 major / 5 minor

Summary. The manuscript introduces randomized quantum optimal control (QOC), in which one optimizes over an ensemble of control waveforms and a probability distribution rather than a single waveform. The central theoretical claim is Theorem 1: there exist a Hamiltonian model, target unitary channel, and tolerance ε>0 such that, under identical control constraints, the minimal time to reach diamond-distance error ≤ε with randomized controls is strictly smaller than with any deterministic control. This is established constructively via an exactly solvable single-qubit model (Eq. 11) in the short-time regime T√(1+f_max^{2})<π with f_max≤1, where the optimal randomized error equals the square of the optimal deterministic error. The SM supplies a complete Pontryagin analysis of the deterministic two-bang optimum, a universal lower bound on mixed-unitary diamond distance, and an explicit two-branch protocol that saturates it. Complementary symmetry constructions (target-preserving twirl and time-reversal pairing) convert a deterministic control into an ensemble that cancels leading coherent error; these are benchmarked on CNOT and match randomized GRAPE. A related boundary-pulse construction is shown to cancel first-order coherent 1-local noise for Clifford targets.

Significance. If the results hold, the paper establishes a genuine resource advantage of randomization inside continuous-time quantum control: under identical amplitude and time constraints, a randomized ensemble can reach a prescribed accuracy faster than any deterministic protocol. The single-qubit existence proof is analytic and essentially complete (PMP structure, candidate comparison, diamond-distance lower bound), and the quadratic relation ε_rand,* = (ε_det,*)^{2} is parameter-free inside the stated regime. The symmetry constructions and R-GRAPE algorithm supply practical tools that reproduce the same quadratic suppression on a two-qubit CNOT benchmark, while the boundary-pulse extension links the framework to noise-robust control. These contributions are of clear interest to the quantum-control and quantum-information communities and open a well-defined research direction.

minor comments (5)
  1. The short-time / f_max≤1 regime that underpins Theorem 1 is stated clearly in the main text and SM, but a single sentence in the main text emphasizing that the quadratic separation is proven only inside this regime (and is not claimed more generally) would help non-specialist readers avoid over-generalization.
  2. Fig. 2(c) would benefit from a brief note in the caption or main text that the faint purple guide is exactly [ε_det_⋆(T)]^{2}, so that the visual match to the symmetry-generated and R-GRAPE curves is immediately interpretable.
  3. The SM error diagnostic (Fig. S1) showing that the dominant deterministic error is Z1X2 is valuable; a one-sentence pointer in the main-text CNOT discussion would make the cancellation mechanism more self-contained for readers who do not immediately consult the SM.
  4. Notation for the averaged unitary U_p,φ and the ensemble-averaged error generator E is introduced cleanly, but a short reminder that global phases are free when comparing unitaries to the target would reduce possible confusion when reading Eqs. (5)–(9).
  5. A few typographical items: “arandomizedQOC” spacing in the abstract/introduction, and occasional missing spaces around math operators in the SM proofs, should be cleaned in production.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: Theorem 1 and the exact quadratic relation are self-contained analytic constructions via PMP and a mixing lower bound; minor self-citations are non-load-bearing.

full rationale

The load-bearing claim (Theorem 1) is proved constructively inside an explicit short-time regime by (i) a complete Pontryagin characterization of the deterministic two-bang optimum (SM Theorem S1 / §§I A–C), (ii) a universal diamond-distance lower bound on any mixed-unitary channel that is proved directly in the SM (Lemma S8 / Prop. S5), and (iii) an explicit two-branch ensemble that saturates the bound, yielding ϵ_rand,* = (ϵ_det,*)^{2} exactly. None of these steps is definitional of the target quantity, none fits a free parameter to data and then “predicts” a related observable, and the regime restriction is stated as a hypothesis rather than hidden. The subsequent symmetry constructions and CNOT numerics are independent supporting evidence. The only self-citations ([35] for the boundary-pulse idea and the in-preparation [40] for a lemma that is re-proved in the SM) are peripheral to the existence proof and do not force the central separation. Hence the derivation chain is self-contained against external benchmarks.

Assumptions & free parameters 3 free parameters · 4 assumptions · 3 invented entities

The central existence claim rests on standard quantum control (Schrödinger equation, diamond norm, PMP) plus the modeling choice that controls are amplitude-bounded and that the single-qubit drift-plus-control Hamiltonian admits a sign-flip symmetry. No free parameters are fitted to obtain the analytic separation; numerical free parameters appear only in the CNOT benchmark. The randomized ensemble itself is the principal invented object.

free parameters (3)
  • f_max (single-qubit example)
    Amplitude bound fixed by the problem statement; not fitted, but the short-time regime and the quadratic relation are stated only for f_max≤1.
  • GRAPE hyperparameters (iterations, learning rate, time discretization, number of random seeds)
    Chosen for the CNOT numerics; best-of-5 seeds reported. Affects the absolute error curves but not the analytic separation.
  • J=1, f_max=10 (two-qubit model)
    Fixed simulation parameters for the CNOT benchmark; not derived from data.
assumptions (4)
  • standard math Diamond distance between unitary channels on a qubit reduces to √(1−|½Tr(U†V)|²) (Prop. S1).
    Used to convert the control problem into maximization of the Y-component; standard result from Watrous.
  • standard math Pontryagin maximum principle supplies necessary conditions for the deterministic single-qubit optimum.
    Invoked throughout SM §I to classify bang-bang and bang-off-bang extremals.
  • domain assumption Sharpened mixing lemma: phase-minimized averaged-unitary error upper-bounds diamond distance of the mixed-unitary channel.
    Cited from Chen et al. (PRX Quantum 2021); used to justify the averaged-unitary picture throughout.
  • domain assumption Conjugation by a symmetry g that preserves the target (up to phase) is realizable by an admissible transformed control waveform.
    Required for the target-preserving twirl construction (Eqs. 16–18); holds for the models considered but is not automatic for arbitrary Hamiltonians.
invented entities (3)
  • Randomized QOC ensemble (waveforms + probabilities)
    purpose: Enlarges the feasible set from single trajectories to convex combinations, enabling coherent-error cancellation.
    Core conceptual object of the paper; no independent experimental evidence outside the constructions given here.
  • Randomized GRAPE (R-GRAPE)
    purpose: Direct gradient optimization of control ensembles and their probabilities.
    Algorithmic generalization introduced in SM Algorithm 1; performance shown only on the CNOT example.
  • Symmetry-generated randomized protocol (twirl + time-reversal pairing)
    purpose: Constructive conversion of a deterministic pulse into an ensemble that cancels leading-order error.
    Explicit four-branch construction for CNOT; independent evidence limited to the numerical match with R-GRAPE.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Randomized Quantum Optimal Control." pith.science (2026). https://pith.science/paper/JHUEFJHF

@misc{pith2026260710946,
  author       = {Pith},
  title        = {Pith review of: Randomized Quantum Optimal Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JHUEFJHF}},
  note         = {Machine review of arXiv:2607.10946}
}
read the original abstract

Quantum optimal control (QOC) aims to find control functions that optimally steer a quantum system toward a target operation. We introduce a \emph{randomized} QOC framework where optimization is carried over an ensemble of control functions and their probabilities, instead of a single set of functions. Using this framework, we prove that randomized QOC can reach a target accuracy faster than any deterministic protocol under the same resource constraints. We also develop general symmetry-based constructions that convert a given control into an ensemble of controls that can systematically reduce the error. We benchmark these constructions for CNOT implementation and find that the resulting randomized protocol quadratically suppresses the error of the optimized deterministic solution. In addition, we introduce randomized GRAPE, which generalizes GRAPE to directly optimize control ensembles and their associated probabilities. Finally, as a related application, we discuss randomized boundary-pulse constructions {that enhance} robustness against coherent noise.

Figures

Figures reproduced from arXiv: 2607.10946 by the authors.

Figure 1
Figure 1. FIG. 1: Schematic comparison between deterministic and [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: Optimal control of CNOT implementation in the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: Noise robustness of CNOT implementation un [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. Randomized product formulas beyond optimal deterministic scaling

    quant-ph 2026-08 accept novelty 8.0 of 10

    Randomized Trotter formulas achieve O(alpha^2) (and O(alpha^3) with a stronger oracle) error scaling for H=A+alpha B with only constant gate overhead, beating proven lower bounds for deterministic formulas.

Reference graph

Works this paper leans on

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

  1. [1]

    As a concrete example, consider any reference wave- form that implementsU targ = CNOT 1→2

    from identity to nontrivial target operations. As a concrete example, consider any reference wave- form that implementsU targ = CNOT 1→2. exactly in the absence of noise. ChoosingB α uniformly from {I, X, Y, Z} ⊗2 gives F V = 1 16 P B B†FV B= 0, so that the first-order coherent-noise contribution is completely removed. Since CNOT is a Clifford gate, eachC...

  2. [2]

    Werschnik and E

    J. Werschnik and E. K. U. Gross, Quantum optimal con- trol theory, Journal of Physics B: Atomic, Molecular and Optical Physics40, R175 (2007)

  3. [3]

    S. J. Glaser, U. Boscain, T. Calarco, C. P. Koch, W. K¨ ockenberger, R. Kosloff, I. Kuprov, B. Luy, S. Schirmer, T. Schulte-Herbr¨ uggen, D. Sugny, and F. K. Wilhelm, Training Schr¨ odinger’s cat: quantum optimal control. Strategic report on current status, visions and goals for research in Europe, European Physical Journal D69, 279 (2015)

  4. [4]

    d’Alessandro,Introduction to quantum control and dy- namics(Chapman and hall/CRC, 2021)

    D. d’Alessandro,Introduction to quantum control and dy- namics(Chapman and hall/CRC, 2021)

  5. [5]

    C. P. Koch, U. Boscain, T. Calarco, G. Dirr, S. Fil- ipp, S. J. Glaser, R. Kosloff, S. Montangero, T. Schulte- Herbr¨ uggen, D. Sugny, and F. K. Wilhelm, Quantum op- timal control in quantum technologies. strategic report on current status, visions and goals for research in eu- rope, EPJ Quantum Technology9, 19 (2022)

  6. [6]

    Ansel, E

    Q. Ansel, E. Dionis, F. Arrouas, B. Peaudecerf, S. Gu´ erin, D. Gu´ ery-Odelin, and D. Sugny, Introduction to theoretical and experimental aspects of quantum op- timal control, Journal of Physics B: Atomic, Molecular and Optical Physics57, 133001 (2024)

  7. [7]

    Khaneja, R

    N. Khaneja, R. Brockett, and S. J. Glaser, Time optimal control in spin systems, Phys. Rev. A63, 032308 (2001)

  8. [8]

    D’Alessandro and M

    D. D’Alessandro and M. Dahleh, Optimal control of two- level quantum systems, IEEE Transactions on Automatic Control46, 866 (2001)

Show all 47 references
  1. [9]

    Boscain, G

    U. Boscain, G. Charlot, J.-P. Gauthier, S. Gu´ erin, and H.-R. Jauslin, Optimal control in laser-induced popula- tion transfer for two- and three-level quantum systems, Journal of Mathematical Physics43, 2107 (2002)

  2. [10]

    Garon, S

    A. Garon, S. J. Glaser, and D. Sugny, Time-optimal control of su(2) quantum operations, Phys. Rev. A88, 043422 (2013)

  3. [11]

    G. C. Hegerfeldt, Driving at the quantum speed limit: Optimal control of a two-level system, Phys. Rev. Lett. 111, 260501 (2013)

  4. [12]

    Boscain, M

    U. Boscain, M. Sigalotti, and D. Sugny, Introduction to the pontryagin maximum principle for quantum optimal control, PRX Quantum2, 030203 (2021)

  5. [13]

    Khaneja, T

    N. Khaneja, T. Reiss, C. Kehlet, T. Schulte-Herbr¨ uggen, and S. J. Glaser, Optimal control of coupled spin dy- namics: design of nmr pulse sequences by gradient as- cent algorithms, Journal of Magnetic Resonance172, 296 (2005)

  6. [14]

    J. P. Palao and R. Kosloff, Optimal control theory for unitary transformations, Phys. Rev. A68, 062308 (2003)

  7. [15]

    Caneva, T

    T. Caneva, T. Calarco, and S. Montangero, Chopped random-basis quantum optimization, Phys. Rev. A84, 022326 (2011)

  8. [16]

    N. Rach, M. M. M¨ uller, T. Calarco, and S. Mon- tangero, Dressing the chopped-random-basis optimiza- tion: A bandwidth-limited access to the trap-free land- scape, Phys. Rev. A92, 062343 (2015)

  9. [17]

    Machnes, E

    S. Machnes, E. Ass´ emat, D. Tannor, and F. K. Wilhelm, Tunable, flexible, and efficient optimization of control pulses for practical qubits, Phys. Rev. Lett.120, 150401 (2018)

  10. [18]

    M. M. M¨ uller, R. S. Said, F. Jelezko, T. Calarco, and S. Montangero, One decade of quantum optimal control in the chopped random basis, Reports on Progress in Physics85, 076001 (2022)

  11. [19]

    Kelly, R

    J. Kelly, R. Barends, B. Campbell, Y. Chen, Z. Chen, B. Chiaro, A. Dunsworth, A. G. Fowler, I.-C. Hoi, E. Jef- frey, A. Megrant, J. Mutus, C. Neill, P. J. J. O’Malley, C. Quintana, P. Roushan, D. Sank, A. Vainsencher, J. Wenner, T. C. White, A. N. Cleland, and J. M. Marti- nis...

  12. [20]

    Dolde, V

    F. Dolde, V. Bergholm, Y. Wang, I. Jakobi, B. Nayde- nov, S. Pezzagna, J. Meijer, F. Jelezko, P. Neumann, T. Schulte-Herbr¨ uggen, J. Biamonte, and J. Wrachtrup, High-fidelity spin entanglement using optimal control, Nature Communications5, 3371 (2014)

  13. [21]

    R. W. Heeres, P. Reinhold, N. Ofek, L. Frunzio, L. Jiang, M. H. Devoret, and R. J. Schoelkopf, Implementing a uni- versal gate set on a logical qubit encoded in an oscillator, Nature Communications8, 94 (2017)

  14. [22]

    Figgatt, A

    C. Figgatt, A. Ostrander, N. M. Linke, K. A. Landsman, D. Zhu, D. Maslov, and C. Monroe, Parallel entangling operations on a universal ion-trap quantum computer, Nature572, 368 (2019)

  15. [23]

    Omran, H

    A. Omran, H. Levine, A. Keesling, G. Semeghini, T. T. Wang, S. Ebadi, H. Bernien, A. S. Zibrov, H. Pichler, S. Choi, J. Cui, M. Rossignolo, P. Rembold, S. Mon- tangero, T. Calarco, M. Endres, M. Greiner, V. Vuleti´ c, and M. D. Lukin, Generation and manipulation of schr¨ oding...

  16. [24]

    Jandura and G

    S. Jandura and G. Pupillo, Time-Optimal Two- and Three-Qubit Gates for Rydberg Atoms, Quantum6, 712 (2022)

  17. [25]

    S. J. Evered, D. Bluvstein, M. Kalinowski, S. Ebadi, T. Manovitz, H. Zhou, S. H. Li, A. A. Geim, T. T. Wang, N. Maskara, H. Levine, G. Semeghini, M. Greiner, V. Vuleti´ c, and M. D. Lukin, High-fidelity parallel entan- gling gates on a neutral-atom quantum computer, Nature 622...

  18. [26]

    Campbell, Random compiler for fast hamiltonian simulation, Physical Review Letters123, 10.1103/phys- revlett.123.070503 (2019)

    E. Campbell, Random compiler for fast hamiltonian simulation, Physical Review Letters123, 10.1103/phys- revlett.123.070503 (2019)

  19. [27]

    A. M. Childs, A. Ostrander, and Y. Su, Faster quantum simulation by randomization, Quantum3, 182 (2019)

  20. [28]

    Boixo, E

    S. Boixo, E. Knill, and R. Somma, Eigenpath traver- sal by phase randomization, Quantum Info. Comput.9, 833–855 (2009)

  21. [29]

    Ouyang, D

    Y. Ouyang, D. R. White, and E. T. Campbell, Compi- lation by stochastic hamiltonian sparsification, Quantum 4, 235 (2020)

  22. [30]

    K. Wan, M. Berta, and E. T. Campbell, Randomized quantum algorithm for statistical phase estimation, Phys. Rev. Lett.129, 030503 (2022)

  23. [31]

    J. M. Martyn and P. Rall, Halving the cost of quantum algorithms with randomization, npj Quantum Informa- tion11, 47 (2025)

  24. [32]

    J. J. Wallman and J. Emerson, Noise tailoring for scalable quantum computation via randomized compiling, Phys. Rev. A94, 052325 (2016)

  25. [33]

    Hashim, R

    A. Hashim, R. K. Naik, A. Morvan, J.-L. Ville, 7 B. Mitchell, J. M. Kreikebaum, M. Davis, E. Smith, C. Iancu, K. P. O’Brien, I. Hincks, J. J. Wallman, J. Emerson, and I. Siddiqi, Randomized compiling for scalable quantum computing on a noisy superconducting quantum processor, ...

  26. [34]

    Viola and E

    L. Viola and E. Knill, Random decoupling schemes for quantum dynamical control and error suppression, Phys. Rev. Lett.94, 060502 (2005)

  27. [35]

    Kern, Randomized dynamical decoupling strategies and improved one-way key rates for quantum cryptogra- phy (2009), arXiv:0906.2927 [quant-ph]

    O. Kern, Randomized dynamical decoupling strategies and improved one-way key rates for quantum cryptogra- phy (2009), arXiv:0906.2927 [quant-ph]

  28. [36]

    C. Yi, L. Kim, and M. Marvian, Faster randomized dy- namical decoupling, Phys. Rev. Lett.136, 010601 (2026)

  29. [37]

    Watrous,The Theory of Quantum Information(Cam- bridge University Press, 2018)

    J. Watrous,The Theory of Quantum Information(Cam- bridge University Press, 2018)

  30. [38]

    Chen, H.-Y

    C.-F. Chen, H.-Y. Huang, R. Kueng, and J. A. Tropp, Concentration for random product formulas, PRX Quan- tum2, 040305 (2021)

  31. [39]

    G. H. Hardy, J. E. Littlewood, and G. P´ olya,Inequalities (Cambridge university press, 1952)

  32. [40]

    J. J. Wallman and S. T. Flammia, Randomized bench- marking with confidence, New Journal of Physics16, 103032 (2014)

  33. [41]

    W. F. Braasch, Jr., L. Kim, and M. Marvian, Limits of Advantage in Randomization in Quantum Protocols (2026), manuscript in preparation

  34. [42]

    M. A. Nielsen, A simple formula for the average gate fi- delity of a quantum dynamical operation, Physics Letters A303, 249–252 (2002). 8 Supplemental Material I. DETERMINISTIC OPTIMUM FOR THE SINGLE-QUBIT EXAMPLE In this section, we prove the deterministic optimal control res...

  35. [43]

    a pure bang-bang control with at most two switchings, f(t)∈ {+f max,−f max},(S71)

  36. [44]

    By Lemma S5, every singular region, every open interval on whichϕ(t) = 0, is an off region with f(t) = 0 almost everywhere

    a bang-off-bang control, f(t) =    s1fmax,0≤t < t 1, 0, t 1 < t < t 2, s2fmax, t 2 < t≤T, 0≤t 1 ≤t 2 ≤T, s 1, s2 ∈ {+1,−1}.(S72) Proof.By the maximum condition in Lemma S4, every region on whichϕ(t)̸= 0 is a bang region, withf(t) =±f max almost everywhere. By Lemma S5, ev...

  37. [45]

    For deterministic GRAPE, every shot uses the same optimized controlf

    Choose a protocol. For deterministic GRAPE, every shot uses the same optimized controlf. For the symmetry- generated randomized protocol, every shot samples uniformly from the four branches in (32). For R-GRAPE, every shot samples from the optimized ensemble according to the o...

  38. [46]

    , N shots, prepareρ in, apply the sampled control branch, and measureO

    For each shots= 1, . . . , N shots, prepareρ in, apply the sampled control branch, and measureO. For simplicity, assumingOis Pauli, the outcome isa s ∈ {±1}

  39. [47]

    S2 shows the resulting observable-estimation errors forO∈ {X 1, Y1, Z1}andN shot = 10000

    Estimate the expectation value ofOby d⟨O⟩= 1 Nshots NshotsX s=1 as.(S126) We then report the observable-estimation error d⟨O⟩ − ⟨O⟩ideal ,⟨O⟩ ideal = Tr OCNOTρinCNOT† .(S127) Fig. S2 shows the resulting observable-estimation errors forO∈ {X 1, Y1, Z1}andN shot = 10000. The mos...

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.