AI-enhanced tuning of quantum dot Hamiltonians toward Majorana modes

Jaros{\l}aw Paw{\l}owski; Mateusz Krawczyk

arxiv: 2601.02149 · v3 · submitted 2026-01-05 · ❄️ cond-mat.mes-hall · cond-mat.dis-nn· cs.AI

AI-enhanced tuning of quantum dot Hamiltonians toward Majorana modes

Mateusz Krawczyk , Jaros{\l}aw Paw{\l}owski This is my paper

Pith reviewed 2026-05-16 17:45 UTC · model grok-4.3

classification ❄️ cond-mat.mes-hall cond-mat.dis-nncs.AI

keywords modesconductancemajoranamapsquantumtowardbroadmodel

0 comments

The pith

A trained vision-transformer network can propose quantum dot parameter updates that produce Majorana zero modes from a wide range of starting conditions.

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

The paper develops a neural network approach to automatically tune quantum dot chains toward hosting Majorana zero modes by analyzing their conductance maps. A deep vision-transformer is trained in an unsupervised way on synthetic data, using a loss function based on expected properties of Majorana modes to learn how parameter changes affect the maps. Once trained, the network can propose updates that, from many different starting detunings, bring the system into a regime with nontrivial zero modes after just one step. Allowing multiple iterations with fresh measurements extends the reachable parameter space significantly.

Core claim

The central claim is that with appropriate training, a deep vision-transformer network can efficiently memorize the relation between Hamiltonian parameters and structures on conductance maps and use it to propose parameter updates for a quantum dot chain that drive the system toward topological phase. Starting from a broad range of initial detunings in parameter space, a single update step is sufficient to generate nontrivial zero modes. Moreover, by enabling an iterative tuning procedure where the system acquires updated conductance maps at each step, the method can address a much larger region of the parameter space.

What carries the argument

Deep vision-transformer network trained unsupervised on synthetic conductance maps with a physics-informed loss incorporating key properties of Majorana zero modes.

Load-bearing premise

The physics-informed loss and synthetic conductance maps generated from the model Hamiltonian accurately represent the signatures and behavior of Majorana modes in real experimental transport measurements, including effects of noise, disorder, and unmodeled physics.

What would settle it

Measure conductance maps from a physical quantum dot chain, feed them to the trained network to get parameter updates, apply those updates to the device, and check if zero-bias peaks appear as predicted.

Figures

Figures reproduced from arXiv: 2601.02149 by Jaros{\l}aw Paw{\l}owski, Mateusz Krawczyk.

**Figure 3.** Figure 3: FIG. 3. Same as in Fig. 2 [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗

**Figure 4.** Figure 4: FIG. 4. Iterative autotuning procedure for (a) model adjust [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗

**Figure 5.** Figure 5: FIG. 5. Three-QD chain Hamiltonian [PITH_FULL_IMAGE:figures/full_fig_p004_5.png] view at source ↗

**Figure 6.** Figure 6: FIG. 6. Conductance maps for the reference parameters [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗

**Figure 7.** Figure 7: FIG. 7. Three-QD chain Hamiltonian [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗

**Figure 8.** Figure 8: FIG. 8. Iterative autotuning procedure for model adjusting [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗

read the original abstract

We propose a neural network-based model capable of learning the broad landscape of working regimes in quantum dot simulators, and using this knowledge to autotune these devices - based on transport measurements - toward obtaining Majorana modes in the structure. The model is trained in an unsupervised manner on synthetic data in the form of conductance maps, using a physics-informed loss that incorporates key properties of Majorana zero modes. We show that, with appropriate training, a deep vision-transformer network can efficiently memorize relation between Hamiltonian parameters and structures on conductance maps and use it to propose parameters update for a quantum dot chain that drive the system toward topological phase. Starting from a broad range of initial detunings in parameter space, a single update step is sufficient to generate nontrivial zero modes. Moreover, by enabling an iterative tuning procedure - where the system acquires updated conductance maps at each step - we demonstrate that the method can address a much larger region of the parameter space.

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

A vision transformer learns to adjust synthetic quantum dot parameters toward Majorana signatures in one step using physics-informed loss on conductance maps, but everything stays inside ideal simulations.

read the letter

The core result is that an unsupervised vision-transformer network, trained on synthetic conductance maps from a quantum dot chain Hamiltonian, can output parameter updates that produce nontrivial zero modes according to the model. A single update often works across a wide range of initial detunings, and an iterative loop extends the reachable parameter space further. This is a straightforward extension of prior ML work on device control, but the specific combination of ViT architecture with a loss that penalizes deviations from zero-bias peaks and gap closing is new for this tuning task. The paper shows the method cleanly in simulation: the network effectively memorizes the relation between parameters and map features and uses it to steer the system into the topological regime without supervised labels. That demonstration is useful as a proof of concept for automated navigation of high-dimensional parameter spaces. The main limitation is that all training and testing data come from the same ideal model Hamiltonian with no disorder, noise, finite-temperature broadening, or lead-coupling variations. Real Majorana searches are dominated by exactly those effects, which can produce similar conductance features without true topological modes. The abstract gives no ablation on loss components, no error bars on the tuning success rate, and no attempt to inject realistic perturbations. Because of this, the claim that the approach will help experimental devices rests on an untested assumption that the synthetic maps are faithful proxies. The work is for people already working on ML for mesoscopic quantum devices who need concrete examples of physics-informed unsupervised tuning. It is coherent on its own terms and shows clear thinking about the problem, so it deserves peer review. Reviewers will likely ask for robustness tests or experimental data, but the simulation results are solid enough to warrant that discussion rather than a desk reject.

Referee Report

3 major / 2 minor

Summary. The manuscript proposes an unsupervised deep vision-transformer network trained on synthetic conductance maps generated from a model Hamiltonian for a quantum dot chain. A physics-informed loss incorporating signatures of Majorana zero modes (zero-bias peak height, gap closing) is used to learn the mapping from conductance features to Hamiltonian parameters. The central claim is that a single gradient-based parameter update suffices to drive the system into the topological regime with nontrivial zero modes from a broad range of initial detunings, with an iterative procedure extending the accessible parameter space.

Significance. If the method generalizes, it offers a practical route to automate tuning of multi-dot devices in high-dimensional parameter spaces, potentially reducing the experimental effort required to locate topological phases. The unsupervised physics-informed training on synthetic data is a methodological strength that avoids the need for labeled experimental datasets.

major comments (3)

[§4] §4 (Results): All demonstrations, including the single-update success from broad detunings, are performed exclusively on clean synthetic conductance maps generated from the ideal model Hamiltonian. No tests with added disorder, finite-temperature broadening, or lead-coupling variations are reported, leaving the generalization to real devices unaddressed.
[§3.2] §3.2 (Loss function): The physics-informed loss combines zero-bias peak height and gap-closing terms, yet no ablation studies quantify the contribution of each term or the sensitivity of the learned mapping to the precise weighting of these components.
[§4.3] §4.3 (Iterative tuning): While iterative updates are shown to enlarge the reachable parameter region, the paper provides no statistics on convergence rate, failure modes, or the number of iterations typically required when starting from random initial detunings.

minor comments (2)

[Figure 2] Figure 2: Axis labels on the conductance maps are small and the color scale lacks explicit units, making quantitative comparison with the loss-function targets difficult.
The abstract states that the network 'memorizes' the parameter-to-map relation; a more precise description of the learned representation (e.g., via attention-map analysis) would strengthen the methodological contribution.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments and the positive assessment of the potential impact of our work. We provide point-by-point responses to the major comments below.

read point-by-point responses

Referee: [§4] §4 (Results): All demonstrations, including the single-update success from broad detunings, are performed exclusively on clean synthetic conductance maps generated from the ideal model Hamiltonian. No tests with added disorder, finite-temperature broadening, or lead-coupling variations are reported, leaving the generalization to real devices unaddressed.

Authors: We acknowledge that the current demonstrations are limited to ideal synthetic data. To address generalization, in the revised manuscript we will include additional results with simulated disorder, finite-temperature effects, and variations in lead coupling. These will demonstrate the robustness of the method and better support its applicability to real devices. revision: yes
Referee: [§3.2] §3.2 (Loss function): The physics-informed loss combines zero-bias peak height and gap-closing terms, yet no ablation studies quantify the contribution of each term or the sensitivity of the learned mapping to the precise weighting of these components.

Authors: We agree that ablation studies would strengthen the paper. We will perform and report ablation experiments in the revised manuscript, including performance metrics when using only one term or varying the weights, to quantify their individual contributions. revision: yes
Referee: [§4.3] §4.3 (Iterative tuning): While iterative updates are shown to enlarge the reachable parameter region, the paper provides no statistics on convergence rate, failure modes, or the number of iterations typically required when starting from random initial detunings.

Authors: We will add statistical analysis of the iterative procedure in the revised version. This will include the distribution of iteration counts for convergence from random initial detunings, success rates, and discussion of observed failure modes. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper trains a vision-transformer network unsupervised on synthetic conductance maps generated from the model Hamiltonian, using a physics-informed loss based on Majorana zero-mode properties to learn a mapping from maps to parameter updates. This is a standard supervised inverse-modeling setup with no reduction of any claimed prediction or result to the inputs by construction. No equations or steps in the abstract or described method exhibit self-definitional equivalence, fitted inputs renamed as predictions, load-bearing self-citations, uniqueness theorems imported from the authors, ansatzes smuggled via citation, or renaming of known results. The derivation remains self-contained as a machine-learning procedure whose outputs are not tautological with the training data generation process.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The method assumes standard properties of Majorana zero modes in one-dimensional topological superconductors and that synthetic data from the target Hamiltonian class suffices for training a generalizable tuner.

axioms (1)

domain assumption Majorana zero modes produce identifiable signatures in differential conductance maps that can be encoded in a physics-informed loss function.
The loss incorporates key properties of Majorana zero modes as stated in the abstract.

pith-pipeline@v0.9.0 · 5471 in / 1284 out tokens · 51309 ms · 2026-05-16T17:45:42.960513+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

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

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

We introduce an unsupervised physics-informed NN-based auto-tuning (PINNAT) model with Majorana physics embedded in the loss function... L(H(P')) = α⟨δP⟩² − M(H(P'))
IndisputableMonolith/Foundation/AlphaCoordinateFixation.lean J_uniquely_calibrated_via_higher_derivative unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

M(H(P)) = p0/2 max[0, 2m0 − Σi>1 mi] with mi = |⟨ψi|M⟩|e−|Ei|/ε and pi quantifying electron-hole symmetry

What do these tags mean?

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

Forward citations

Cited by 1 Pith paper

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

Information in Many-body Eigenstates: A Question of Learnability
quant-ph 2026-05 unverdicted novelty 6.0

Machine learning reconstruction accuracy is substantially higher for spectral-edge eigenstates than for mid-spectrum eigenstates, providing a new quantitative measure of information content in many-body quantum states.

Reference graph

Works this paper leans on

56 extracted references · 56 canonical work pages · cited by 1 Pith paper · 2 internal anchors

[1]

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)

work page 2012
[2]

Beenakker, Search for Majorana fermions in supercon- ductors, Annual Review of Condensed Matter Physics4, 113 (2013)

C. Beenakker, Search for Majorana fermions in supercon- ductors, Annual Review of Condensed Matter Physics4, 113 (2013)

work page 2013
[3]

J. D. Sau, R. M. Lutchyn, S. Tewari, and S. Das Sarma, Generic new platform for topological quantum compu- tation using semiconductor heterostructures, Phys. Rev. Lett.104, 040502 (2010)

work page 2010
[4]

Y. Oreg, G. Refael, and F. von Oppen, Helical liquids and Majorana bound states in quantum wires, Phys. Rev. Lett.105, 177002 (2010)

work page 2010
[5]

Leijnse and K

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

work page 2012
[6]

R. M. Lutchyn, E. P. A. M. Bakkers, L. P. Kouwen- hoven, P. Krogstrup, C. M. Marcus, and Y. Oreg, Majo- rana zero modes in superconductor–semiconductor het- erostructures, Nature Reviews Materials3, 52 (2018)

work page 2018
[7]

Leumer, M

N. Leumer, M. Grifoni, B. Muralidharan, and M. Mar- ganska, Linear and nonlinear transport across a finite Ki- taev chain: An exact analytical study, Phys. Rev. B103, 165432 (2021)

work page 2021
[8]

A. Y. Kitaev, Unpaired Majorana fermions in quantum wires, Physics-Uspekhi44, s131 (2001), translation from: Usp. Fiz. Nauk 171 (2001) 131–136, supplement№10

work page 2001
[9]

Valentini, F

M. Valentini, F. Pe˜ naranda, A. Hofmann, M. Brauns, R. Hauschild, P. Krogstrup, P. San-Jose, E. Prada, R. Aguado, and G. Katsaros, Nontopological zero-bias peaks in full-shell nanowires induced by flux-tunable an- dreev states, Science373, 82 (2021)

work page 2021
[10]

Sasaki, S

S. Sasaki, S. D. Franceschi, J. M. Elzerman, W. G. van der Wiel, M. Eto, S. Tarucha, and L. P. Kouwen- hoven, Kondo effect in an integer-spin quantum dot, Na- ture405, 764 (2000)

work page 2000
[11]

D. I. Pikulin, J. P. Dahlhaus, M. Wimmer, H. Schome- rus, and C. W. J. Beenakker, A zero-voltage conductance peak from weak antilocalization in a Majorana nanowire, New Journal of Physics14, 125011 (2012)

work page 2012
[12]

Kells, D

G. Kells, D. Meidan, and P. W. Brouwer, Low-energy subgap states in multichannelp-wave superconducting wires, Phys. Rev. B85, 060507 (2012)

work page 2012
[13]

I. C. Fulga, A. Haim, A. R. Akhmerov, and Y. Oreg, Adaptive tuning of Majorana fermions in a quantum dot chain, New Journal of Physics15, 045020 (2013)

work page 2013
[14]

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)

work page 2012
[15]

J. D. Sau and S. D. Sarma, Realizing a robust practical Majorana chain in a quantum-dot-superconductor linear array, Nature Communications3, 964 (2012)

work page 2012
[16]

T. Dvir, G. Wang, N. van Loo, C.-X. Liu, G. P. Mazur, A. Bordin, S. L. D. ten Haaf, J.-Y. Wang, D. van Driel, F. Zatelli, X. Li, F. K. Malinowski, S. Gazibegovic, G. Badawy, E. P. A. M. Bakkers, M. Wimmer, and L. P. Kouwenhoven, Realization of a minimal Kitaev chain in coupled quantum dots, Nature614, 445 (2023)

work page 2023
[17]

Bordin, C.-X

A. Bordin, C.-X. Liu, T. Dvir, F. Zatelli, S. L. D. ten Haaf, D. van Driel, G. Wang, N. van Loo, Y. Zhang, J. C. Wolff, T. V. Caekenberghe, G. Badawy, S. Gazibegovic, E. P. A. M. Bakkers, M. Wimmer, L. P. Kouwenhoven, and G. P. Mazur, Enhanced Majorana stability in a three- site Kitaev chain, Nature Nanotechnology20, 726 (2025)

work page 2025
[18]

Mohseni, H

M. Mohseni, H. Allami, D. Miravet, D. J. Gayowsky, M. Korkusinski, and P. Hawrylak, Majorana excitons in a Kitaev chain of semiconductor quantum dots in a nanowire, Nanomaterials13, 10.3390/nano13162293 (2023)

work page doi:10.3390/nano13162293 2023
[19]

Ezawa, Even-odd effect on robustness of Majorana edge states in short Kitaev chains, Phys

M. Ezawa, Even-odd effect on robustness of Majorana edge states in short Kitaev chains, Phys. Rev. B109, 11 TABLE II. Comparison of reference and PINNAT parameter sets for the three-dot chain. Energies are given in meV. In the presented sets ∆ n does not change withn: ∆ n = ∆. Configuration⟨µ n⟩V Z ∆t 1 t2 λ1 λ2 VZ maxn tn VZ√ ∆2+⟨µn⟩2 ⟨λn⟩ arctan ⟨µn ⟩ ∆...

work page 2024
[20]

Borsoi, N

F. Borsoi, N. W. Hendrickx, V. John, M. Meyer, S. Motz, F. van Riggelen, A. Sammak, S. L. de Snoo, G. Scap- pucci, and M. Veldhorst, Shared control of a 16 semicon- ductor quantum dot crossbar array, Nature Nanotechnol- ogy19, 21 (2024)

work page 2024
[21]

A. R. Mills, D. M. Zajac, M. J. Gullans, F. J. Schupp, T. M. Hazard, and J. R. Petta, Shuttling a single charge across a one-dimensional array of silicon quantum dots, Nature Communications10, 1063 (2019)

work page 2019
[22]

Shandilya, S

A. Shandilya, S. Kapila, R. Krishnan, B. Weber, and B. Muralidharan, Unified simulation framework for experimentally-observed spin-valley locking in MoS2 quantum dots: Implications for qubit applications, ACS Applied Nano Materials8, 14949 (2025)

work page 2025
[23]

Darulov´ a, S

J. Darulov´ a, S. Pauka, N. Wiebe, K. Chan, G. Gardener, M. Manfra, M. Cassidy, and M. Troyer, Autonomous tun- ing and charge-state detection of gate-defined quantum dots, Phys. Rev. Appl.13, 054005 (2020)

work page 2020
[24]

Lunczer, P

L. Lunczer, P. Leubner, M. Endres, V. L. M¨ uller, C. Br¨ une, H. Buhmann, and L. W. Molenkamp, Ap- proaching quantization in macroscopic quantum spin hall devices through gate training, Phys. Rev. Lett.123, 047701 (2019)

work page 2019
[25]

J. P. Zwolak and J. M. Taylor, Colloquium: Advances in automation of quantum dot devices control, Rev. Mod. Phys.95, 011006 (2023)

work page 2023
[26]

M.-A. Roux, J. Rivard, V. Yon, A. Morel, D. Leclerc, C. Rohrbacher, E. B. Ndiaye, F. F. Tafuri, B. Bono, S. Kubicek, R. Loo, Y. Shimura, J. Jussot, C. God- frin, D. Wan, K. D. Greve, M.-A. T´ etrault, D. Drouin, C. Lupien, M. Pioro-Ladri` ere, and E. Dupont-Ferrier, Rapid autotuning of a sige quantum dot into the single-electron regime with machine learni...

work page internal anchor Pith review Pith/arXiv arXiv 2025
[27]

M. P. R. Losert, D. Denora, B. van Straaten, M. Chan, S. D. Oosterhout, L. Stehouwer, G. Scappucci, M. Veld- horst, and J. P. Zwolak, Automated electrostatic char- acterization of quantum dot devices in single- and bi- layer heterostructures (2025), arXiv:2601.00067 [cond- mat.mes-hall]

work page arXiv 2025
[28]

Donarini and M

A. Donarini and M. Grifoni,Quantum Transport in In- teracting Nanojunctions(Springer, 2024)

work page 2024
[29]

G. E. Blonder, M. Tinkham, and T. M. Klapwijk, Tran- sition from metallic to tunneling regimes in supercon- ducting microconstrictions: Excess current, charge im- balance, and supercurrent conversion, Phys. Rev. B25, 4515 (1982)

work page 1982
[30]

R. Koch, D. van Driel, A. Bordin, J. L. Lado, and E. Gre- plova, Adversarial hamiltonian learning of quantum dots in a minimal Kitaev chain, Phys. Rev. Appl.20, 044081 (2023)

work page 2023
[31]

J. Wang, S. Paesani, R. Santagati, S. Knauer, A. A. Gen- tile, N. Wiebe, M. Petruzzella, J. L. O’Brien, J. G. Rarity, A. Laing, and M. G. Thompson, Experimental quantum hamiltonian learning, Nature Physics13, 551 (2017)

work page 2017
[32]

Gebhart, R

V. Gebhart, R. Santagati, A. A. Gentile, E. M. Gauger, D. Craig, N. Ares, L. Banchi, F. Marquardt, L. Pezz` e, and C. Bonato, Learning quantum systems, Nature Re- views Physics5, 141 (2023)

work page 2023
[33]

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)

work page 2024
[34]

J. R. Taylor, J. D. Sau, and S. Das Sarma, Machine learn- ing the disorder landscape of Majorana nanowires, Phys. Rev. Lett.132, 206602 (2024)

work page 2024
[35]

J. R. Taylor and S. Das Sarma, Neural network based deep learning analysis of semiconductor quantum dot qubits for automated control, Phys. Rev. B111, 035301 (2025)

work page 2025
[36]

van Driel, R

D. van Driel, R. Koch, V. P. M. Sietses, S. L. D. ten Haaf, C.-X. Liu, F. Zatelli, B. Roovers, A. Bordin, N. van Loo, G. Wang, J. C. Wolff, G. P. Mazur, T. Dvir, I. Kulesh, Q. Wang, A. M. Bozkurt, S. Gazibegovic, G. Badawy, E. P. A. M. Bakkers, M. Wimmer, S. Goswami, J. L. Lado, L. P. Kouwenhoven, and E. Greplova, Cross- platform autonomous control of min...

work page arXiv 2024
[37]

J. R. Taylor and S. Das Sarma, Mitigating dis- order and optimizing topological indicators with vision-transformer-based neural networks in Majorana nanowires, Phys. Rev. B112, L041110 (2025)

work page 2025
[38]

Raissi, P

M. Raissi, P. Perdikaris, and G. Karniadakis, Physics- informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics378, 686 (2019)

work page 2019
[39]

Kovachki, Z

N. Kovachki, Z. Li, B. Liu, K. Azizzadenesheli, K. Bhat- tacharya, A. Stuart, and A. Anandkumar, Neural opera- tor: Learning maps between function spaces with appli- cations to PDEs, Journal of Machine Learning Research 24, 1 (2023)

work page 2023
[40]

Dosovitskiy, L

A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, An im- age is worth 16x16 words: Transformers for image recog- nition at scale, inInternational Conference on Learning Representations(2021)

work page 2021
[41]

Sticlet, C

D. Sticlet, C. Bena, and P. Simon, Spin and Majorana polarization in topological superconducting wires, Phys. Rev. Lett.108, 096802 (2012). 12

work page 2012
[42]

M. M. Ma´ ska and T. Doma´ nski, Polarization of the Ma- jorana quasiparticles in the Rashba chain, Scientific Re- ports7, 16193 (2017)

work page 2017
[43]

Szumniak, D

P. Szumniak, D. Chevallier, D. Loss, and J. Klinovaja, Spin and charge signatures of topological superconduc- tivity in Rashba nanowires, Phys. Rev. B96, 041401 (2017)

work page 2017
[44]

Stanescu,Introduction to Topological Quantum Matter & Quantum Computation(CRC Press, 2024)

T. Stanescu,Introduction to Topological Quantum Matter & Quantum Computation(CRC Press, 2024)

work page 2024
[45]

C.-X. Liu, J. D. Sau, T. D. Stanescu, and S. Das Sarma, Andreev bound states versus majorana bound states in quantum dot-nanowire-superconductor hybrid struc- tures: Trivial versus topological zero-bias conductance peaks, Phys. Rev. B96, 075161 (2017)

work page 2017
[46]

Prada, P

E. Prada, P. San-Jose, M. W. A. de Moor, A. Geresdi, E. J. H. 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)

work page 2020
[47]

M. M. Bronstein, J. Bruna, T. Cohen, and P. Veliˇ ckovi´ c, Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges, arXiv preprint arXiv:2104.13478 (2021)

work page internal anchor Pith review Pith/arXiv arXiv 2021
[48]

Paw lowski and M

J. Paw lowski and M. Krawczyk, Identification of quan- tum entanglement with siamese convolutional neural net- works and semisupervised learning, Phys. Rev. Appl.22, 014068 (2024)

work page 2024
[49]

Krawczyk, J

M. Krawczyk, J. Paw lowski, M. M. Ma´ ska, and K. Roszak, Data-driven criteria for quantum correlations, Phys. Rev. A109, 022405 (2024)

work page 2024
[50]

Norambuena, M

A. Norambuena, M. Mattheakis, F. J. Gonz´ alez, and R. Coto, Physics-informed neural networks for quantum control, Phys. Rev. Lett.132, 010801 (2024)

work page 2024
[51]

Krawczyk, P

M. Krawczyk, P. Bal´ aˇ z, K. Roszak, and J. Paw lowski, Learning quantum tomography from incomplete mea- surements (2025), arXiv:2506.19428 [quant-ph]

work page arXiv 2025
[52]

Paw lowski and M

J. Paw lowski and M. Krawczyk, Learning hamil- tonians for solid-state quantum simulators (2026), arXiv:2603.02889 [cond-mat.mes-hall]

work page arXiv 2026
[53]

Christiansen and M

T. Christiansen and M. Zworski, A mathematical for- mulation of the Mahaux–Weidenm¨ uller formula for the scattering matrix, Journal of Physics A: Mathematical and Theoretical42, 415202 (2009)

work page 2009
[54]

Bordin, F

A. Bordin, F. J. Bennebroek Evertsz’, B. Roovers, J. D. Torres 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. Seoane Souto, N. van Loo, and L. P. Kouwenhoven, Probing majorana localization of a phase- controlled three-site Kitaev chain with an additional quantum dot, Nature Communications17, ...

work page 2026
[55]

W. Wang, Z. Dang, Y. Hu, P. Fua, and M. Salz- mann, Backpropagation-friendly eigendecomposition, in Proceedings of the 33rd International Conference on Neu- ral Information Processing Systems(Curran Associates Inc., Red Hook, NY, USA, 2019)

work page 2019
[56]

T. Li, Z. Shi, J. Zhao, and M. Lin, Amortized eigen- decomposition for neural networks, inAdvances in Neu- ral Information Processing Systems, Vol. 37, edited by A. Globerson, L. Mackey, D. Belgrave, A. Fan, U. Pa- quet, J. Tomczak, and C. Zhang (Curran Associates, Inc.,

work page

[1] [1]

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)

work page 2012

[2] [2]

Beenakker, Search for Majorana fermions in supercon- ductors, Annual Review of Condensed Matter Physics4, 113 (2013)

C. Beenakker, Search for Majorana fermions in supercon- ductors, Annual Review of Condensed Matter Physics4, 113 (2013)

work page 2013

[3] [3]

J. D. Sau, R. M. Lutchyn, S. Tewari, and S. Das Sarma, Generic new platform for topological quantum compu- tation using semiconductor heterostructures, Phys. Rev. Lett.104, 040502 (2010)

work page 2010

[4] [4]

Y. Oreg, G. Refael, and F. von Oppen, Helical liquids and Majorana bound states in quantum wires, Phys. Rev. Lett.105, 177002 (2010)

work page 2010

[5] [5]

Leijnse and K

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

work page 2012

[6] [6]

R. M. Lutchyn, E. P. A. M. Bakkers, L. P. Kouwen- hoven, P. Krogstrup, C. M. Marcus, and Y. Oreg, Majo- rana zero modes in superconductor–semiconductor het- erostructures, Nature Reviews Materials3, 52 (2018)

work page 2018

[7] [7]

Leumer, M

N. Leumer, M. Grifoni, B. Muralidharan, and M. Mar- ganska, Linear and nonlinear transport across a finite Ki- taev chain: An exact analytical study, Phys. Rev. B103, 165432 (2021)

work page 2021

[8] [8]

A. Y. Kitaev, Unpaired Majorana fermions in quantum wires, Physics-Uspekhi44, s131 (2001), translation from: Usp. Fiz. Nauk 171 (2001) 131–136, supplement№10

work page 2001

[9] [9]

Valentini, F

M. Valentini, F. Pe˜ naranda, A. Hofmann, M. Brauns, R. Hauschild, P. Krogstrup, P. San-Jose, E. Prada, R. Aguado, and G. Katsaros, Nontopological zero-bias peaks in full-shell nanowires induced by flux-tunable an- dreev states, Science373, 82 (2021)

work page 2021

[10] [10]

Sasaki, S

S. Sasaki, S. D. Franceschi, J. M. Elzerman, W. G. van der Wiel, M. Eto, S. Tarucha, and L. P. Kouwen- hoven, Kondo effect in an integer-spin quantum dot, Na- ture405, 764 (2000)

work page 2000

[11] [11]

D. I. Pikulin, J. P. Dahlhaus, M. Wimmer, H. Schome- rus, and C. W. J. Beenakker, A zero-voltage conductance peak from weak antilocalization in a Majorana nanowire, New Journal of Physics14, 125011 (2012)

work page 2012

[12] [12]

Kells, D

G. Kells, D. Meidan, and P. W. Brouwer, Low-energy subgap states in multichannelp-wave superconducting wires, Phys. Rev. B85, 060507 (2012)

work page 2012

[13] [13]

I. C. Fulga, A. Haim, A. R. Akhmerov, and Y. Oreg, Adaptive tuning of Majorana fermions in a quantum dot chain, New Journal of Physics15, 045020 (2013)

work page 2013

[14] [14]

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)

work page 2012

[15] [15]

J. D. Sau and S. D. Sarma, Realizing a robust practical Majorana chain in a quantum-dot-superconductor linear array, Nature Communications3, 964 (2012)

work page 2012

[16] [16]

T. Dvir, G. Wang, N. van Loo, C.-X. Liu, G. P. Mazur, A. Bordin, S. L. D. ten Haaf, J.-Y. Wang, D. van Driel, F. Zatelli, X. Li, F. K. Malinowski, S. Gazibegovic, G. Badawy, E. P. A. M. Bakkers, M. Wimmer, and L. P. Kouwenhoven, Realization of a minimal Kitaev chain in coupled quantum dots, Nature614, 445 (2023)

work page 2023

[17] [17]

Bordin, C.-X

A. Bordin, C.-X. Liu, T. Dvir, F. Zatelli, S. L. D. ten Haaf, D. van Driel, G. Wang, N. van Loo, Y. Zhang, J. C. Wolff, T. V. Caekenberghe, G. Badawy, S. Gazibegovic, E. P. A. M. Bakkers, M. Wimmer, L. P. Kouwenhoven, and G. P. Mazur, Enhanced Majorana stability in a three- site Kitaev chain, Nature Nanotechnology20, 726 (2025)

work page 2025

[18] [18]

Mohseni, H

M. Mohseni, H. Allami, D. Miravet, D. J. Gayowsky, M. Korkusinski, and P. Hawrylak, Majorana excitons in a Kitaev chain of semiconductor quantum dots in a nanowire, Nanomaterials13, 10.3390/nano13162293 (2023)

work page doi:10.3390/nano13162293 2023

[19] [19]

Ezawa, Even-odd effect on robustness of Majorana edge states in short Kitaev chains, Phys

M. Ezawa, Even-odd effect on robustness of Majorana edge states in short Kitaev chains, Phys. Rev. B109, 11 TABLE II. Comparison of reference and PINNAT parameter sets for the three-dot chain. Energies are given in meV. In the presented sets ∆ n does not change withn: ∆ n = ∆. Configuration⟨µ n⟩V Z ∆t 1 t2 λ1 λ2 VZ maxn tn VZ√ ∆2+⟨µn⟩2 ⟨λn⟩ arctan ⟨µn ⟩ ∆...

work page 2024

[20] [20]

Borsoi, N

F. Borsoi, N. W. Hendrickx, V. John, M. Meyer, S. Motz, F. van Riggelen, A. Sammak, S. L. de Snoo, G. Scap- pucci, and M. Veldhorst, Shared control of a 16 semicon- ductor quantum dot crossbar array, Nature Nanotechnol- ogy19, 21 (2024)

work page 2024

[21] [21]

A. R. Mills, D. M. Zajac, M. J. Gullans, F. J. Schupp, T. M. Hazard, and J. R. Petta, Shuttling a single charge across a one-dimensional array of silicon quantum dots, Nature Communications10, 1063 (2019)

work page 2019

[22] [22]

Shandilya, S

A. Shandilya, S. Kapila, R. Krishnan, B. Weber, and B. Muralidharan, Unified simulation framework for experimentally-observed spin-valley locking in MoS2 quantum dots: Implications for qubit applications, ACS Applied Nano Materials8, 14949 (2025)

work page 2025

[23] [23]

Darulov´ a, S

J. Darulov´ a, S. Pauka, N. Wiebe, K. Chan, G. Gardener, M. Manfra, M. Cassidy, and M. Troyer, Autonomous tun- ing and charge-state detection of gate-defined quantum dots, Phys. Rev. Appl.13, 054005 (2020)

work page 2020

[24] [24]

Lunczer, P

L. Lunczer, P. Leubner, M. Endres, V. L. M¨ uller, C. Br¨ une, H. Buhmann, and L. W. Molenkamp, Ap- proaching quantization in macroscopic quantum spin hall devices through gate training, Phys. Rev. Lett.123, 047701 (2019)

work page 2019

[25] [25]

J. P. Zwolak and J. M. Taylor, Colloquium: Advances in automation of quantum dot devices control, Rev. Mod. Phys.95, 011006 (2023)

work page 2023

[26] [26]

M.-A. Roux, J. Rivard, V. Yon, A. Morel, D. Leclerc, C. Rohrbacher, E. B. Ndiaye, F. F. Tafuri, B. Bono, S. Kubicek, R. Loo, Y. Shimura, J. Jussot, C. God- frin, D. Wan, K. D. Greve, M.-A. T´ etrault, D. Drouin, C. Lupien, M. Pioro-Ladri` ere, and E. Dupont-Ferrier, Rapid autotuning of a sige quantum dot into the single-electron regime with machine learni...

work page internal anchor Pith review Pith/arXiv arXiv 2025

[27] [27]

M. P. R. Losert, D. Denora, B. van Straaten, M. Chan, S. D. Oosterhout, L. Stehouwer, G. Scappucci, M. Veld- horst, and J. P. Zwolak, Automated electrostatic char- acterization of quantum dot devices in single- and bi- layer heterostructures (2025), arXiv:2601.00067 [cond- mat.mes-hall]

work page arXiv 2025

[28] [28]

Donarini and M

A. Donarini and M. Grifoni,Quantum Transport in In- teracting Nanojunctions(Springer, 2024)

work page 2024

[29] [29]

G. E. Blonder, M. Tinkham, and T. M. Klapwijk, Tran- sition from metallic to tunneling regimes in supercon- ducting microconstrictions: Excess current, charge im- balance, and supercurrent conversion, Phys. Rev. B25, 4515 (1982)

work page 1982

[30] [30]

R. Koch, D. van Driel, A. Bordin, J. L. Lado, and E. Gre- plova, Adversarial hamiltonian learning of quantum dots in a minimal Kitaev chain, Phys. Rev. Appl.20, 044081 (2023)

work page 2023

[31] [31]

J. Wang, S. Paesani, R. Santagati, S. Knauer, A. A. Gen- tile, N. Wiebe, M. Petruzzella, J. L. O’Brien, J. G. Rarity, A. Laing, and M. G. Thompson, Experimental quantum hamiltonian learning, Nature Physics13, 551 (2017)

work page 2017

[32] [32]

Gebhart, R

V. Gebhart, R. Santagati, A. A. Gentile, E. M. Gauger, D. Craig, N. Ares, L. Banchi, F. Marquardt, L. Pezz` e, and C. Bonato, Learning quantum systems, Nature Re- views Physics5, 141 (2023)

work page 2023

[33] [33]

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)

work page 2024

[34] [34]

J. R. Taylor, J. D. Sau, and S. Das Sarma, Machine learn- ing the disorder landscape of Majorana nanowires, Phys. Rev. Lett.132, 206602 (2024)

work page 2024

[35] [35]

J. R. Taylor and S. Das Sarma, Neural network based deep learning analysis of semiconductor quantum dot qubits for automated control, Phys. Rev. B111, 035301 (2025)

work page 2025

[36] [36]

van Driel, R

D. van Driel, R. Koch, V. P. M. Sietses, S. L. D. ten Haaf, C.-X. Liu, F. Zatelli, B. Roovers, A. Bordin, N. van Loo, G. Wang, J. C. Wolff, G. P. Mazur, T. Dvir, I. Kulesh, Q. Wang, A. M. Bozkurt, S. Gazibegovic, G. Badawy, E. P. A. M. Bakkers, M. Wimmer, S. Goswami, J. L. Lado, L. P. Kouwenhoven, and E. Greplova, Cross- platform autonomous control of min...

work page arXiv 2024

[37] [37]

J. R. Taylor and S. Das Sarma, Mitigating dis- order and optimizing topological indicators with vision-transformer-based neural networks in Majorana nanowires, Phys. Rev. B112, L041110 (2025)

work page 2025

[38] [38]

Raissi, P

M. Raissi, P. Perdikaris, and G. Karniadakis, Physics- informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations, Journal of Computational Physics378, 686 (2019)

work page 2019

[39] [39]

Kovachki, Z

N. Kovachki, Z. Li, B. Liu, K. Azizzadenesheli, K. Bhat- tacharya, A. Stuart, and A. Anandkumar, Neural opera- tor: Learning maps between function spaces with appli- cations to PDEs, Journal of Machine Learning Research 24, 1 (2023)

work page 2023

[40] [40]

Dosovitskiy, L

A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly, J. Uszkoreit, and N. Houlsby, An im- age is worth 16x16 words: Transformers for image recog- nition at scale, inInternational Conference on Learning Representations(2021)

work page 2021

[41] [41]

Sticlet, C

D. Sticlet, C. Bena, and P. Simon, Spin and Majorana polarization in topological superconducting wires, Phys. Rev. Lett.108, 096802 (2012). 12

work page 2012

[42] [42]

M. M. Ma´ ska and T. Doma´ nski, Polarization of the Ma- jorana quasiparticles in the Rashba chain, Scientific Re- ports7, 16193 (2017)

work page 2017

[43] [43]

Szumniak, D

P. Szumniak, D. Chevallier, D. Loss, and J. Klinovaja, Spin and charge signatures of topological superconduc- tivity in Rashba nanowires, Phys. Rev. B96, 041401 (2017)

work page 2017

[44] [44]

Stanescu,Introduction to Topological Quantum Matter & Quantum Computation(CRC Press, 2024)

T. Stanescu,Introduction to Topological Quantum Matter & Quantum Computation(CRC Press, 2024)

work page 2024

[45] [45]

C.-X. Liu, J. D. Sau, T. D. Stanescu, and S. Das Sarma, Andreev bound states versus majorana bound states in quantum dot-nanowire-superconductor hybrid struc- tures: Trivial versus topological zero-bias conductance peaks, Phys. Rev. B96, 075161 (2017)

work page 2017

[46] [46]

Prada, P

E. Prada, P. San-Jose, M. W. A. de Moor, A. Geresdi, E. J. H. 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)

work page 2020

[47] [47]

M. M. Bronstein, J. Bruna, T. Cohen, and P. Veliˇ ckovi´ c, Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges, arXiv preprint arXiv:2104.13478 (2021)

work page internal anchor Pith review Pith/arXiv arXiv 2021

[48] [48]

Paw lowski and M

J. Paw lowski and M. Krawczyk, Identification of quan- tum entanglement with siamese convolutional neural net- works and semisupervised learning, Phys. Rev. Appl.22, 014068 (2024)

work page 2024

[49] [49]

Krawczyk, J

M. Krawczyk, J. Paw lowski, M. M. Ma´ ska, and K. Roszak, Data-driven criteria for quantum correlations, Phys. Rev. A109, 022405 (2024)

work page 2024

[50] [50]

Norambuena, M

A. Norambuena, M. Mattheakis, F. J. Gonz´ alez, and R. Coto, Physics-informed neural networks for quantum control, Phys. Rev. Lett.132, 010801 (2024)

work page 2024

[51] [51]

Krawczyk, P

M. Krawczyk, P. Bal´ aˇ z, K. Roszak, and J. Paw lowski, Learning quantum tomography from incomplete mea- surements (2025), arXiv:2506.19428 [quant-ph]

work page arXiv 2025

[52] [52]

Paw lowski and M

J. Paw lowski and M. Krawczyk, Learning hamil- tonians for solid-state quantum simulators (2026), arXiv:2603.02889 [cond-mat.mes-hall]

work page arXiv 2026

[53] [53]

Christiansen and M

T. Christiansen and M. Zworski, A mathematical for- mulation of the Mahaux–Weidenm¨ uller formula for the scattering matrix, Journal of Physics A: Mathematical and Theoretical42, 415202 (2009)

work page 2009

[54] [54]

Bordin, F

A. Bordin, F. J. Bennebroek Evertsz’, B. Roovers, J. D. Torres 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. Seoane Souto, N. van Loo, and L. P. Kouwenhoven, Probing majorana localization of a phase- controlled three-site Kitaev chain with an additional quantum dot, Nature Communications17, ...

work page 2026

[55] [55]

W. Wang, Z. Dang, Y. Hu, P. Fua, and M. Salz- mann, Backpropagation-friendly eigendecomposition, in Proceedings of the 33rd International Conference on Neu- ral Information Processing Systems(Curran Associates Inc., Red Hook, NY, USA, 2019)

work page 2019

[56] [56]

T. Li, Z. Shi, J. Zhao, and M. Lin, Amortized eigen- decomposition for neural networks, inAdvances in Neu- ral Information Processing Systems, Vol. 37, edited by A. Globerson, L. Mackey, D. Belgrave, A. Fan, U. Pa- quet, J. Tomczak, and C. Zhang (Curran Associates, Inc.,

work page