Pith. sign in

REVIEW 2 major objections 5 minor 59 references

Noise resilience itself can be the control signal that finds protected Majorana states.

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 →

T0 review · deepseek-v4-flash

2026-08-04 00:18 UTC pith:XY6CM7VA

load-bearing objection Using the noise-resilience of E0 as a CMA-ES loss is a genuinely different tuning idea; the numerics are solid, but the experimental-relevance claim is the soft spot. the 2 major comments →

arxiv 2511.01531 v1 pith:XY6CM7VA submitted 2025-11-03 cond-mat.mes-hall

Machine-learned tuning to protected states by probing noise resilience

classification cond-mat.mes-hall
keywords Majorana bound statesKitaev chainsquantum dotsCMA-ESnoise resiliencesweet spotsprotected statesautonomous tuning
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.

This paper aims to establish a general principle: when a protected quantum state exists only at a hard-to-find point in parameter space, you can locate it by deliberately perturbing the system and steering toward the configuration least sensitive to those perturbations. Concretely, the authors simulate quantum-dot-based Kitaev chains, inject random fluctuations into the dot levels, and use an evolutionary optimizer (CMA-ES) to minimize the average ground-state energy splitting. Across two- to five-site chains, the optimization converges to sweet spots with nearly zero splitting, well-localized Majorana bound states at the ends, and a finite excitation gap. The method needs only the ground-state splitting as feedback, which is in principle measurable in real devices, and the authors argue the strategy extends to any protected state, not just Majoranas.

Core claim

The central claim is that tuning to a Majorana sweet spot can be achieved by minimizing the energy splitting of a Kitaev chain subjected to local noise. Using CMAA-ES with a loss function given by the average ground-state splitting over random onsite-potential fluctuations, the simulations converge to configurations with near-zero splitting, Majorana polarization close to one at the outer dots, and a finite excitation gap. The correspondence between noise resilience and protection is verified for chain lengths two through five, asymmetric arrays, and interacting systems.

What carries the argument

The key object is the loss function L(ε) = (1/P) Σ_p E0(ε + η^(p)), the ground-state splitting averaged over P random realizations of local level fluctuations η. This quantity turns the abstract idea of 'protection' into a concrete scalar to minimize; the optimizer then walks the gate voltages toward the most noise-resilient configuration. The paper supplements this with two diagnostics: the Majorana polarization at the outer dots, which measures whether the zero-energy states are localized at the ends, and the excitation gap, which ensures the sweet spot is separated from higher states.

Load-bearing premise

The protocol assumes that the ground-state splitting E0 used by the optimizer is the same quantity an experimentalist can extract from the device accurately and quickly enough, but the paper never models that measurement layer.

What would settle it

Tune a real three-dot Kitaev device with this protocol using a measured E0 (e.g., conductance or Ramsey), then independently verify the final configuration via nonlocal conductance or parity readout; if a low noise-averaged E0 does not come with a near-zero actual splitting and end-localized states separated by a finite gap, the noise-averaged loss is not a sufficient tuning signal.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • An experiment could run this tuning loop using only a measurement of the ground-state splitting E0, e.g., via conductance spectroscopy or Ramsey interferometry, without resolving the Majorana polarization directly.
  • The protocol automatically handles high-dimensional parameter spaces and multiple sweet spots: in the two-site chain, different search bounds lead to different but equally protected configurations.
  • The amplitude and distribution of the injected fluctuations act as a dial: widening fluctuations on the outer dots trades a slightly lower Majorana polarization for a larger excitation gap.
  • The method is not specific to Majorana systems; any protected state whose defining feature is robustness to local perturbations could be located by the same noise-minimization strategy.
  • Longer chains require extra care: the excitation gap no longer converges to a unique value, suggesting that adding a gap term to the loss function would make the protocol more reliable as the system grows.

Where Pith is reading between the lines

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

  • The success of the method suggests a broader 'noise-as-guide' design principle: rather than fighting noise, one can deliberately inject it and use the system's response to navigate to robust operating points, potentially applicable to spin qubits, superconducting circuits, or other tunable quantum devices.
  • Because the loss function uses only the average |E0| over fluctuations, the method is likely insensitive to the specific noise spectrum, but real devices have non-stationary and correlated noise; testing the protocol under colored noise or with a biased E0 measurement is a natural next step.
  • The optimizer found the same sweet spot in all 50 runs for a given search bound, hinting that the loss landscape is smooth enough that simpler optimizers might also work and that the landscape may have symmetries that aid convergence.
  • The trade-off between Majorana polarization and gap controlled by the fluctuation weighting suggests that future implementations could tailor the injected noise distribution to emphasize whichever property matters more for a given experiment.

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

2 major / 5 minor

Summary. The paper proposes a machine-learning protocol for tuning to protected states in quantum-dot-based Kitaev chains. The method injects random fluctuations into the onsite quantum-dot potentials and uses CMA-ES to minimize the noise-averaged ground-state energy splitting E0 (Eq. (4)). Simulations for 2-, 3-, 4-, and 5-site chains show that, from random initial potentials, the algorithm converges to configurations with near-zero E0, high Majorana polarization at the outer dots, and a finite excitation gap for the shorter chains. The authors test robustness against electron-electron interactions, finite Zeeman fields, and asymmetric couplings, and argue the strategy is general and readily applicable to experiments because only E0 is required.

Significance. The core idea — using the noise resilience of a protected state as the tuning metric — is appealing and likely to interest the quantum-dot Majorana community. The numerical evidence is solid where complete: 50 independent runs per system with reported medians and spreads, and final configurations are checked with quantities (Majorana polarization, excitation gap) not present in the loss, which partially answers circularity concerns. The demonstration that a simple E0-based loss finds sweet spots in 2- and 3-site chains is a useful step beyond previous conductance-based ML tuning protocols. The paper also tests interactions, asymmetry, and varying chain length. Its main limitations are the absence of a measurement-resolution model in the experimental claim and the uncontrolled excitation gap in longer chains, which are discussed below.

major comments (2)
  1. [Conclusions and Eq. (4)] The loss in Eq. (4) is evaluated with exact BdG ground-state splittings. The statement that the method is 'readily applied to experiments' requires that E0 can be measured with sufficient accuracy and speed. A real measurement has finite resolution (thermal broadening, integration time, readout noise); if measured E0 saturates at zero for all configurations below a threshold, CMA-ES loses the relative loss information it needs. No model of this measurement layer is included. I recommend adding a simulation where L uses a resolution-limited or noisy E0 (e.g., max(E0,δ)+noise) and showing that convergence survives, or deriving a quantitative requirement on δ relative to W and the sweet-spot basin.
  2. [Longer Kitaev chains and Fig. 4] Eq. (4) does not control the excitation gap, although a protected state requires a finite gap. For 4- and 5-site chains the final gap varies substantially across runs (purple shades in Fig. 4), and the text acknowledges: 'The excitation gap also does not always converge to the same values... A possible way of overcoming these limitations could be to include direct measurements of the excitation gap into the loss function.' This admitted limitation means the central claim of reliable tuning to protected states is not established for longer chains. The loss should be augmented to include the gap, or the conclusions should be restricted to the 2- and 3-site cases where the gap is consistently finite.
minor comments (5)
  1. [References] References [46] and [47] are identical (same arXiv:2504.13702). They should be merged or one removed.
  2. [Fig. 2] The text refers to the 'right blue cross' and 'left blue cross' in Fig. 2(a,b), but the crosses are not separately labeled in the figure. Adding explicit labels would improve clarity.
  3. [Eq. (4)] The superscript i is used both for the population member index and for the site index in the definition of the loss function. Using k for the population member would avoid confusion.
  4. [General methods] The experimental overhead of the optimization loop (P=200 noise samples per candidate, 20 candidates per generation) is not discussed. A rough time estimate or a note on parallelization would help assess practical feasibility.
  5. [Abstract] The abstract claims the method applies 'including but not limited to isolated Majorana bound states,' but no non-Majorana protected state is demonstrated. This is acceptable as a forward-looking remark, but it should be phrased as a conjecture rather than a demonstrated result.

Circularity Check

1 steps flagged

Loss directly targets the defining sweet-spot property (robust E0), but independent MP and gap checks keep the central claim from being purely circular.

specific steps
  1. self definitional [Tuning Kitaev chains to protected states; Eq. (4) and Fig. 2(d)]
    "MBS sweet spots are characterized by three properties: (i) a ground state degeneracy, i.e., E0 = 0 ... (iii) highly localized MBSs on opposite ends ... A characteristic and defining feature of these sweet spots is the robustness of E0 against parameter fluctuations. ... Each configuration is then assessed via a loss function based on a numerical evaluation of E0, L(ε(i)) = 1/P Σ_p E0(ε(i)+η(p)) (4)."

    The loss minimized by CMA-ES is literally the fluctuation-averaged ground-state splitting L = (1/P)Σ E0(ε+η). Since Eq. (4) is built from the same E0 whose robustness is called the defining feature of a sweet spot, finding low-L configurations is largely guaranteed by construction to land at points with robust E0. That part of the claimed 'tuning to protected states' is definitional rather than predictive. The claim is rescued from full circularity by the independent monitors not appearing in the loss — Majorana polarization ≈ 1 and finite excitation gap Eex — which show that the robust-E0 points found are genuine MBS sweet spots rather than generic near-degeneracies.

full rationale

The core numerical demonstration is self-contained: Eq. (4) is evaluated from exact BdG E0 in the same model used to define sweet spots, and the convergence statistics in Figs. 2–4 come from fresh simulations, not from external fitted values. The main circular element is definitional: the loss is the noise-averaged E0, and the paper explicitly states that robustness of E0 is a defining feature of a sweet spot, so minimizing that loss aims directly at the defining property. This would be fully circular if the paper claimed only 'low loss implies protected', but it also verifies quantities not in the loss — Majorana polarization and excitation gap — over many runs and chain lengths, giving independent content. Citations to the authors' prior work [39,51] provide the model and sweet-spot phenomenology but are not used as a uniqueness theorem or as the sole justification for convergence; the numerical tests stand on their own. The acknowledged measurement limitations (thermal broadening, Ramsey readout) affect experimental feasibility but are not a circularity argument. Score 4 reflects the partial self-definitional character of the loss function, balanced by independent MBS-quality checks.

Axiom & Free-Parameter Ledger

6 free parameters · 5 axioms · 0 invented entities

The central results depend on several hand-tuned algorithmic and model parameters (W, β, bounds, npop, P, termination threshold) and on domain assumptions about the noise model and the experimental measurability of E0. No new physical entities are introduced; the contribution is a control strategy.

free parameters (6)
  • Noise amplitude W = 0.075Δ, 0.1Δ, and 0.125Δ
    Hand-chosen per chain length; footnote [57] says W controls the trade-off between E0 and MP and the final E0 value.
  • Outer-dot noise weight β = 1 and 2
    Chosen for 3-site chains to steer convergence toward sweet spots with larger excitation gap at the cost of reduced MP.
  • CMA-ES population size npop and noise samples P = npop=20, P=200
    Algorithmic hyperparameters; no convergence or sensitivity analysis is provided.
  • Initial distribution and search bounds = Δ≤ε1,3≤2Δ and 0.2Δ≤ε2≤1.2Δ (or negative ε2 range)
    Mimic experimental variations, but the converged sweet spot (left vs right) depends on these bounds, as stated in the text.
  • Termination threshold = Not given in main text
    Defined in Supplementary Material Sec. A; unavailable in the reviewed text.
  • Microscopic model parameters = t_i=0.5Δ, t_so_i=0.2Δ, V_z,i=1.5Δ (odd), V_z,i=0 (even)
    Fixed parameters; the authors claim qualitative independence but all simulations are run at this point.
axioms (5)
  • domain assumption The QD array is described by the BdG Hamiltonian Eqs. (1)-(2) in the non-interacting limit; interactions are neglected in the main text.
    Main text around Eq. (1) states interactions are neglected and 'verified' not to change results in SM Sec. C.
  • domain assumption Random uncorrelated uniform fluctuations η_i ∈ [−W,W] of the QD levels model the relevant local noise in experiments.
    Introduced in Eq. (4) and the surrounding text; no experimental noise model is given.
  • domain assumption Minimizing the average E0 loss with CMA-ES converges to MBS sweet spots rather than trivial accidental zero-energy crossings.
    The optimization landscape is not analyzed; the claim is supported only by the numerical MP checks.
  • domain assumption Majorana polarization (Eq. 3) is a reliable local marker of MBS localization.
    Adopted from Refs. [39,53-56] as the verification metric.
  • domain assumption E0 is measurable in experiments with sufficient fidelity to drive the optimization (conductance spectroscopy or Ramsey experiments).
    States in the Conclusions; no measurement-error model is included.

pith-pipeline@v1.3.0-alltime-deepseek · 10797 in / 12544 out tokens · 137033 ms · 2026-08-04T00:18:29.921507+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of Machine-learned tuning to protected states by probing noise resilience." pith.science (2026). https://pith.science/paper/XY6CM7VA

@misc{pith2026251101531,
  author       = {Pith},
  title        = {Pith review of: Machine-learned tuning to protected states by probing noise resilience},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XY6CM7VA}},
  note         = {Machine review of arXiv:2511.01531}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Protected states are promising for quantum technologies due to their intrinsic resilience against noise. However, such states often emerge at discrete points or small regions in parameter space and are thus difficult to find in experiments. In this work, we present a machine-learning method for tuning to protected regimes, based on injecting noise into the system and searching directly for the most noise-resilient configuration. We illustrate this method by considering short quantum dot-based Kitaev chains which we subject to random parameter fluctuations. Using the covariance matrix adaptation evolutionary strategy we minimize the typical resulting ground state splitting, which makes the system converge to a protected configuration with well-separated Majorana bound states. We verify the robustness of our method by considering finite Zeeman fields, electron-electron repulsion, asymmetric couplings, and varying the length of the Kitaev chain. Our work provides a reliable method for tuning to protected states, including but not limited to isolated Majorana bound states.

Figures

Figures reproduced from arXiv: 2511.01531 by Jacob Benestad, Jeroen Danon, Martin Leijnse, Nicol\'as Mart\'inez-Valero, Rodrigo A. Dourado, Rub\'en Seoane Souto.

Figure 1
Figure 1. Figure 1: FIG. 1. Tuning to MBS sweet spots via parameter fluctua [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Results for the 2-site Kitaev chain. (a,b) Energy [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Results for longer Kitaev chains. Evolution of the [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Results for the 3-site Kitaev chain. (a) Sketch of the [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

discussion (0)

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

Reference graph

Works this paper leans on

59 extracted references · 7 linked inside Pith

  1. [1]

    D. A. Lidar and K. Birgitta Whaley, Decoherence-free subspaces and subsystems, inIrreversible Quantum Dy- namics, edited by F. Benatti and R. Floreanini (Springer Berlin Heidelberg, 2003) pp. 83–120

  2. [2]

    Gyenis, A

    A. Gyenis, A. Di Paolo, J. Koch, A. Blais, A. A. Houck, and D. I. Schuster, Moving beyond the transmon: Noise- protected superconducting quantum circuits, PRX Quan- tum2, 030101 (2021)

  3. [3]

    Danon, A

    J. Danon, A. Chatterjee, A. Gyenis, and F. Kuem- meth, Protected solid-state qubits, Appl. Phys. Lett. 119, 260502 (2021)

  4. [4]

    D. Vion, A. Aassime, A. Cottet, P. Joyez, H. Pothier, C. Urbina, D. Esteve, and M. H. Devoret, Manipulating the quantum state of an electrical circuit, Science296, 886 (2002)

  5. [5]

    Ithier, E

    G. Ithier, E. Collin, P. Joyez, P. J. Meeson, D. Vion, D. Esteve, F. Chiarello, A. Shnirman, Y. Makhlin, J. Schriefl, and G. Sch¨ on, Decoherence in a supercon- ducting quantum bit circuit, Phys. Rev. B72, 134519 (2005)

  6. [6]

    J. Koch, T. M. Yu, J. Gambetta, A. A. Houck, D. I. Schuster, J. Majer, A. Blais, M. H. Devoret, S. M. Girvin, and R. J. Schoelkopf, Charge-insensitive qubit design de- rived from the cooper pair box, Phys. Rev. A76, 042319 (2007)

  7. [7]

    Martins, F

    F. Martins, F. K. Malinowski, P. D. Nissen, E. Barnes, S. Fallahi, G. C. Gardner, M. J. Manfra, C. M. Mar- cus, and F. Kuemmeth, Noise suppression using symmet- ric exchange gates in spin qubits, Phys. Rev. Lett.116, 116801 (2016)

  8. [8]

    M. D. Reed, B. M. Maune, R. W. Andrews, M. G. Borselli, K. Eng, M. P. Jura, A. A. Kiselev, T. D. Ladd, S. T. Merkel, I. Milosavljevic, E. J. Pritchett, M. T. Rakher, R. S. Ross, A. E. Schmitz, A. Smith, J. A. Wright, M. F. Gyure, and A. T. Hunter, Reduced sensitivity to charge noise in semiconductor spin qubits via symmetric operation, Phys. Rev. Lett.116...

  9. [9]

    E. A. Sete, M. J. Reagor, N. Didier, and C. T. Rigetti, Charge- and flux-insensitive tunable superconducting qubit, Phys. Rev. Appl.8, 024004 (2017)

  10. [10]

    L. B. Nguyen, Y.-H. Lin, A. Somoroff, R. Mencia, N. Grabon, and V. E. Manucharyan, High-coherence flux- onium qubit, Phys. Rev. X9, 041041 (2019)

  11. [11]

    Didier, E

    N. Didier, E. A. Sete, J. Combes, and M. P. da Silva, ac flux sweet spots in parametrically modulated supercon- ducting qubits, Phys. Rev. Appl.12, 054015 (2019)

  12. [12]

    Frees, S

    A. Frees, S. Mehl, J. K. Gamble, M. Friesen, and S. N. Coppersmith, Adiabatic two-qubit gates in capacitively coupled quantum dot hybrid qubits, npj Quantum Inf.5, 73 (2019)

  13. [13]

    Huang, P

    Z. Huang, P. S. Mundada, A. Gyenis, D. I. Schuster, A. A. Houck, and J. Koch, Engineering dynamical sweet spots to protect qubits from 1/fnoise, Phys. Rev. Appl. 15, 034065 (2021)

  14. [14]

    Didier, Flux control of superconducting qubits at dy- namical sweet spots, arXiv:1912.09416 (2019)

    N. Didier, Flux control of superconducting qubits at dy- namical sweet spots, arXiv:1912.09416 (2019)

  15. [15]

    S. S. Hong, A. T. Papageorge, P. Sivarajah, G. Crossman, N. Didier, A. M. Polloreno, E. A. Sete, S. W. Turkowski, M. P. da Silva, and B. R. Johnson, Demonstration of a parametrically activated entangling gate protected from flux noise, Phys. Rev. A101, 012302 (2020)

  16. [16]

    Nayak, S

    C. Nayak, S. H. Simon, A. Stern, M. Freedman, and S. Das Sarma, Non-abelian anyons and topological quan- tum computation, Rev. Mod. Phys.80, 1083 (2008)

  17. [17]

    A. Y. Kitaev, Unpaired Majorana fermions in quantum wires, Physics-Uspekhi44, 131 (2001)

  18. [18]

    A. Y. Kitaev, Fault-tolerant quantum computation by anyons, Annals of physics303, 2 (2003)

  19. [19]

    Leijnse and K

    M. Leijnse and K. Flensberg, Introduction to topological superconductivity and Majorana fermions, Semiconduc- tor Science and Technology27, 124003 (2012)

  20. [20]

    Alicea, New directions in the pursuit of Majorana fermions in solid state systems, Reports on progress in physics75, 076501 (2012)

    J. Alicea, New directions in the pursuit of Majorana fermions in solid state systems, Reports on progress in physics75, 076501 (2012)

  21. [21]

    Das Sarma, M

    S. Das Sarma, M. Freedman, and C. Nayak, Majorana zero modes and topological quantum computation, npj Quantum Inf.1, 15001 (2015)

  22. [22]

    Aguado, Majorana quasiparticles in condensed mat- ter, La Rivista del Nuovo Cimento40, 523 (2017)

    R. Aguado, Majorana quasiparticles in condensed mat- ter, La Rivista del Nuovo Cimento40, 523 (2017)

  23. [23]

    Flensberg, F

    K. Flensberg, F. von Oppen, and A. Stern, Engineered platforms for topological superconductivity and Ma- jorana zero modes, Nature Reviews Materials6, 944 (2021)

  24. [24]

    Prada, P

    E. Prada, P. San-Jose, M. W. de Moor, A. Geresdi, E. J. Lee, J. Klinovaja, D. Loss, J. Nyg ˚ ard, R. Aguado, and L. P. Kouwenhoven, From Andreev to Majorana bound states in hybrid superconductor–semiconductor nanowires, Nature Reviews Physics2, 575 (2020)

  25. [25]

    C. W. J. Beenakker, Search for non-Abelian Majorana braiding statistics in superconductors, SciPost Phys. Lect. Notes , 15 (2020)

  26. [26]

    Das Sarma, In search of Majorana, Nature Physics19, 165 (2023)

    S. Das Sarma, In search of Majorana, Nature Physics19, 165 (2023)

  27. [27]

    L. Kouwenhoven, Perspective on Majorana bound-states in hybrid superconductor-semiconductor nanowires, Modern Physics Letters B39, 2540002 (2025), https://doi.org/10.1142/S0217984925400020. 6

  28. [28]

    Seoane Souto and R

    R. Seoane Souto and R. Aguado, Subgap states in semiconductor-superconductor devices for quantum tech- nologies: Andreev qubits and minimal Majorana chains, inNew Trends and Platforms for Quantum Technolo- gies, edited by R. Aguado, R. Citro, M. Lewenstein, and M. Stern (Springer Nature Switzerland, Cham, 2024) pp. 133–223

  29. [29]

    J. D. Sau and S. D. Sarma, Realizing a robust practical majorana chain in a quantum-dot-superconductor linear array, Nature communications3, 964 (2012)

  30. [30]

    Leijnse and K

    M. Leijnse and K. Flensberg, Parity qubits and poor man’s Majorana bound states in double quantum dots, Phys. Rev. B86, 134528 (2012)

  31. [31]

    C.-X. Liu, G. Wang, T. Dvir, and M. Wimmer, Tun- able superconducting coupling of quantum dots via Andreev bound states in semiconductor-superconductor nanowires, Phys. Rev. Lett.129, 267701 (2022)

  32. [32]

    Aghaee, A

    M. Aghaee, A. Akkala, Z. Alam, R. Ali, A. Al- caraz Ramirez, M. Andrzejczuk, A. E. Antipov, P. Aseev, M. Astafev, B. Bauer, J. Becker, S. Boddapati, F. Boekhout, J. Bommer, T. Bosma, L. Bourdet, S. Boutin, P. Caroff, L. Casparis, M. Cassidy, S. Cha- toor, A. W. Christensen, N. Clay, W. S. Cole, F. Corsetti, A. Cui, P. Dalampiras, A. Dokania, G. de Lange, ...

  33. [33]

    Thamm and B

    M. Thamm and B. Rosenow, Machine learning optimiza- tion of Majorana hybrid nanowires, Phys. Rev. Lett.130, 116202 (2023)

  34. [34]

    Thamm and B

    M. Thamm and B. Rosenow, Conductance based machine learning of optimal gate voltages for disordered Majorana wires, Phys. Rev. B109, 045132 (2024)

  35. [35]

    Benestad, A

    J. Benestad, A. Tsintzis, R. S. Souto, M. Leijnse, E. van Nieuwenburg, and J. Danon, Machine-learned tuning of artificial Kitaev chains from tunneling spectroscopy mea- surements, Phys. Rev. B110, 075402 (2024)

  36. [36]

    van Driel, R

    D. van Driel, R. Koch, V. P. Sietses, S. L. ten Haaf, C.- X. Liu, F. Zatelli, B. Roovers, A. Bordin, N. van Loo, G. Wang,et al., Cross-platform autonomous control of minimal kitaev chains, arXiv preprint arXiv:2405.04596 (2024)

  37. [37]

    C. J. Bolech and E. Demler, Observing Majorana bound states inp-wave superconductors using noise measure- ments in tunneling experiments, Phys. Rev. Lett.98, 237002 (2007)

  38. [38]

    T. O. Rosdahl, A. Vuik, M. Kjaergaard, and A. R. Akhmerov, Andreev rectifier: A nonlocal conductance signature of topological phase transitions, Phys. Rev. B 97, 045421 (2018)

  39. [39]

    Tsintzis, R

    A. Tsintzis, R. S. Souto, and M. Leijnse, Creating and de- tecting poor man’s Majorana bound states in interacting quantum dots, Physical Review B106, L201404 (2022)

  40. [40]

    R. S. Souto, A. Tsintzis, M. Leijnse, and J. Danon, Probing Majorana localization in minimal Kitaev chains through a quantum dot, Phys. Rev. Res.5, 043182 (2023)

  41. [41]

    R. A. Dourado, J. Danon, M. Leijnse, and R. S. Souto, Measuring coherence factors of states in superconductors through local current (2025), arXiv:2507.20696 [cond- mat.mes-hall]

  42. [42]

    T. Dvir, G. Wang, N. van Loo, C.-X. Liu, G. P. Mazur, A. Bordin, S. L. Ten Haaf, J.-Y. Wang, D. van Driel, F. Zatelli,et al., Realization of a minimal kitaev chain in coupled quantum dots, Nature614, 445 (2023)

  43. [43]

    S. L. Ten Haaf, Q. Wang, A. M. Bozkurt, C.-X. Liu, I. Kulesh, P. Kim, D. Xiao, C. Thomas, M. J. Man- fra, T. Dvir,et al., A two-site Kitaev chain in a two- dimensional electron gas, Nature630, 329 (2024)

  44. [44]

    S. L. Ten Haaf, Y. Zhang, Q. Wang, A. Bordin, C.-X. Liu, I. Kulesh, V. P. Sietses, C. G. Prosko, D. Xiao, C. Thomas,et al., Observation of edge and bulk states in a three-site Kitaev chain, Nature641, 890–895 (2025)

  45. [45]

    Bordin, C.-X

    A. Bordin, C.-X. Liu, T. Dvir, F. Zatelli, S. L. Ten Haaf, D. van Driel, G. Wang, N. Van Loo, Y. Zhang, J. C. Wolff,et al., Enhanced majorana stability in a three-site kitaev chain, Nature Nanotechnology , 1 (2025)

  46. [47]

    Bordin, F

    A. Bordin, F. J. B. Evertsz’, B. Roovers, J. D. T. Luna, W. D. Huisman, F. Zatelli, G. P. Mazur, S. L. D. ten Haaf, G. Badawy, E. P. A. M. Bakkers, C.-X. Liu, R. S. Souto, N. van Loo, and L. P. Kouwenhoven, Probing Majorana localization of a phase-controlled three-site Kitaev chain with an additional quantum dot (2025), arXiv:2504.13702

  47. [48]

    van Loo, F

    N. van Loo, F. Zatelli, G. O. Steffensen, B. Roovers, G. Wang, T. V. Caekenberghe, A. Bordin, D. van Driel, Y. Zhang, W. D. Huisman, G. Badawy, E. P. A. M. Bakkers, G. P. Mazur, R. Aguado, and L. P. Kouwen- hoven, Single-shot parity readout of a minimal Kitaev chain (2025), arXiv:2507.01606 [cond-mat.mes-hall]

  48. [49]

    Zhang, I

    Y. Zhang, I. Kulesh, S. L. D. ten Haaf, N. van Loo, F. Za- telli, T. Degroote, C. G. Prosko, and S. Goswami, Gate reflectometry in a minimal Kitaev chain device (2025), arXiv:2508.06403 [cond-mat.mes-hall]

  49. [50]

    Luethi, H

    M. Luethi, H. F. Legg, D. Loss, and J. Klinovaja, From perfect to imperfect poor man’s majoranas in minimal kitaev chains, Phys. Rev. B110, 245412 (2024)

  50. [51]

    R. A. Dourado, M. Leijnse, and R. S. Souto, Majorana 7 sweet spots in three-site Kitaev chains, Phys. Rev. B111, 235409 (2025)

  51. [52]

    Hansen, The cma evolution strategy: A tutorial (2023), arXiv:1604.00772 [cs.LG]

    N. Hansen, The cma evolution strategy: A tutorial (2023), arXiv:1604.00772 [cs.LG]

  52. [53]

    Sticlet, C

    D. Sticlet, C. Bena, and P. Simon, Spin and Majorana polarization in topological superconducting wires, Phys- ical Review Letters108, 096802 (2012)

  53. [54]

    Sedlmayr and C

    N. Sedlmayr and C. Bena, Visualizing Majorana bound states in one and two dimensions using the generalized Majorana polarization, Physical Review B92, 115115 (2015)

  54. [55]

    Sedlmayr, J

    N. Sedlmayr, J. Aguiar-Hualde, and C. Bena, Majorana bound states in open quasi-one-dimensional and two- dimensional systems with transverse Rashba coupling, Physical Review B93, 155425 (2016)

  55. [56]

    O. A. Awoga and J. Cayao, Identifying trivial and Majo- rana zero-energy modes using the Majorana polarization, Phys. Rev. B110, 165404 (2024)

  56. [57]

    This value is roughly half of the excitation gap for 2-site Kitaev chains [39]. The amplitudeWcontrols the opti- mization ofE 0 and the MP, with larger values penalizing MBS overlaps more heavily, thereby increasing the MP, but also increasing the final value ofE 0

  57. [58]

    C.-X. Liu, S. Miles, A. Bordin, S. L. D. ten Haaf, G. P. Mazur, A. M. Bozkurt, and M. Wimmer, Scaling up a sign-ordered Kitaev chain without magnetic flux control, Phys. Rev. Res.7, L012045 (2025)

  58. [59]

    Tsintzis, R

    A. Tsintzis, R. S. Souto, K. Flensberg, J. Danon, and M. Leijnse, Majorana qubits and non-abelian physics in quantum dot–based minimal Kitaev chains, PRX Quan- tum5, 010323 (2024)

  59. [60]

    H. Pan, S. Das Sarma, and C.-X. Liu, Rabi and Ram- sey oscillations of a Majorana qubit in a quantum dot- superconductor array, Phys. Rev. B111, 075416 (2025)