pith. sign in

arxiv: 2606.29281 · v1 · pith:TXRH4BFBnew · submitted 2026-06-28 · ❄️ cond-mat.mes-hall · quant-ph

Learning Inhomogeneous Heisenberg Hamiltonians in Nanographene Spin Chains

Greta Lupi , Saketh Ravuri , Chenxiao Zhao , Weidan Zhang , Cesare Roncaglia , Renxiang Liu , Xinliang Feng , Daniele Passerone
show 4 more authors
Pascal Ruffieux Roman Fasel Jose L. Lado Gon\c{c}alo Catarina
This is my paper

Pith reviewed 2026-06-30 02:32 UTC · model grok-4.3

classification ❄️ cond-mat.mes-hall quant-ph
keywords machine learningHeisenberg Hamiltoniannanographenespin chainsscanning tunneling spectroscopyinhomogeneous exchangequantum spin systems
0
0 comments X

The pith

Machine learning reconstructs spatially modulated exchange interactions in nanographene spin chains directly from inelastic STS maps.

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

The paper introduces a local machine learning method that infers spatially varying exchange couplings in spin chains from inelastic scanning tunneling spectroscopy data without depending on chain length. This targets the frequent assumption of uniform interactions in low-dimensional magnets even when structural or environmental factors create local differences. The approach identifies uniform and inhomogeneous cases in nanographene systems, produces models that match measured spectra, and reproduces the observed dependence of the excitation gap on system size. A sympathetic reader would see this as a practical bridge between local measurements and effective many-body descriptions of quantum spin materials.

Core claim

We leverage a local, length-independent machine learning methodology to reconstruct spatially modulated exchange interactions directly from inelastic scanning tunneling spectroscopy maps. We demonstrate this approach with nanographene spin chains, identifying both near-uniform and inhomogeneous regimes across the synthesized magnets. The reconstructed models quantitatively reproduce the experimental spectra and recover the correct scaling of the excitation gap with system size. Our results establish a general strategy to bridge local spectroscopic measurements with effective many-body Hamiltonians.

What carries the argument

A local, length-independent machine learning methodology that takes inelastic STS maps as input and outputs spatially modulated exchange couplings in an effective Heisenberg Hamiltonian.

If this is right

  • The method distinguishes near-uniform from inhomogeneous exchange regimes in synthesized nanographene magnets.
  • Reconstructed Hamiltonians match experimental inelastic spectra at a quantitative level.
  • The models recover the correct dependence of the excitation gap on chain length.
  • The workflow provides a general route from local spectroscopic data to effective many-body spin Hamiltonians.

Where Pith is reading between the lines

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

  • The same reconstruction pipeline could be tested on other low-dimensional spin systems such as molecular chains or trapped-ion arrays where local coupling variations are suspected.
  • If additional interaction terms like anisotropy or longer-range couplings prove necessary, the method could be extended by enlarging the parameter space while keeping the local fitting structure.
  • Quantitative agreement on gap scaling suggests the approach may help predict behavior in devices or larger assemblies built from the same nanographene building blocks.

Load-bearing premise

The measured inelastic STS maps arise from an effective Heisenberg spin Hamiltonian whose only spatial variation occurs in the exchange couplings, with no major contributions from other terms or from the measurement process.

What would settle it

Apply the reconstructed Hamiltonians to chains longer than those used in training and check whether the predicted excitation gap continues to follow the observed scaling; mismatch in gap values or spectra would falsify the reconstruction.

Figures

Figures reproduced from arXiv: 2606.29281 by Cesare Roncaglia, Chenxiao Zhao, Daniele Passerone, Gon\c{c}alo Catarina, Greta Lupi, Jose L. Lado, Pascal Ruffieux, Renxiang Liu, Roman Fasel, Saketh Ravuri, Weidan Zhang, Xinliang Feng.

Figure 1
Figure 1. Figure 1: FIG. 1. (a) Illustration of an olympicene spin chain on [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. (a) Experimental d [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. (a) Spin density distribution of a single Olympicene [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. On-surface synthesis route used to fabricate the [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p008_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Reconstruction of the exchange couplings for (a) an [PITH_FULL_IMAGE:figures/full_fig_p009_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: FIG. 9. Two representative [PITH_FULL_IMAGE:figures/full_fig_p010_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: FIG. 10. Experimental [PITH_FULL_IMAGE:figures/full_fig_p011_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: FIG. 11. (a) Training and validation loss curves and (b) pre [PITH_FULL_IMAGE:figures/full_fig_p012_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: FIG. 12. Fidelity as a function of the relative noise strength [PITH_FULL_IMAGE:figures/full_fig_p013_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: FIG. 13 [PITH_FULL_IMAGE:figures/full_fig_p014_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: FIG. 14 [PITH_FULL_IMAGE:figures/full_fig_p015_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: FIG. 15 [PITH_FULL_IMAGE:figures/full_fig_p016_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: FIG. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: FIG. 17 [PITH_FULL_IMAGE:figures/full_fig_p017_17.png] view at source ↗
Figure 19
Figure 19. Figure 19: FIG. 19 [PITH_FULL_IMAGE:figures/full_fig_p017_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: FIG. 20 [PITH_FULL_IMAGE:figures/full_fig_p018_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: FIG. 21 [PITH_FULL_IMAGE:figures/full_fig_p018_21.png] view at source ↗
Figure 23
Figure 23. Figure 23: FIG. 23 [PITH_FULL_IMAGE:figures/full_fig_p019_23.png] view at source ↗
Figure 25
Figure 25. Figure 25: FIG. 25 [PITH_FULL_IMAGE:figures/full_fig_p019_25.png] view at source ↗
Figure 24
Figure 24. Figure 24: FIG. 24 [PITH_FULL_IMAGE:figures/full_fig_p019_24.png] view at source ↗
Figure 26
Figure 26. Figure 26: FIG. 26 [PITH_FULL_IMAGE:figures/full_fig_p019_26.png] view at source ↗
Figure 27
Figure 27. Figure 27: FIG. 27 [PITH_FULL_IMAGE:figures/full_fig_p020_27.png] view at source ↗
Figure 29
Figure 29. Figure 29: FIG. 29 [PITH_FULL_IMAGE:figures/full_fig_p020_29.png] view at source ↗
Figure 28
Figure 28. Figure 28: FIG. 28 [PITH_FULL_IMAGE:figures/full_fig_p020_28.png] view at source ↗
Figure 30
Figure 30. Figure 30: FIG. 30 [PITH_FULL_IMAGE:figures/full_fig_p021_30.png] view at source ↗
Figure 32
Figure 32. Figure 32: FIG. 32 [PITH_FULL_IMAGE:figures/full_fig_p021_32.png] view at source ↗
Figure 33
Figure 33. Figure 33: FIG. 33 [PITH_FULL_IMAGE:figures/full_fig_p022_33.png] view at source ↗
Figure 34
Figure 34. Figure 34: FIG. 34 [PITH_FULL_IMAGE:figures/full_fig_p022_34.png] view at source ↗
read the original abstract

Inferring microscopic Hamiltonians from experimental data is a central challenge in quantum materials and quantum simulation. In low-dimensional spin systems, exchange interactions are often assumed to be spatially uniform, despite structural and environmental inhomogeneities that can locally modify the coupling. Here, we leverage a local, length-independent machine learning methodology to reconstruct spatially modulated exchange interactions directly from inelastic scanning tunneling spectroscopy maps. We demonstrate this approach with nanographene spin chains, identifying both near-uniform and inhomogeneous regimes across the synthesized magnets. The reconstructed models quantitatively reproduce the experimental spectra and recover the correct scaling of the excitation gap with system size. Our results establish a general strategy to bridge local spectroscopic measurements with effective many-body Hamiltonians.

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.

Referee Report

3 major / 2 minor

Summary. The manuscript presents a local, length-independent machine learning methodology to reconstruct spatially modulated exchange interactions J(r) directly from inelastic scanning tunneling spectroscopy (STS) maps in nanographene spin chains. It applies the approach to both near-uniform and inhomogeneous regimes, claiming that the resulting Heisenberg models quantitatively reproduce the experimental spectra and recover the correct scaling of the excitation gap with system size.

Significance. If the central claims hold after detailed validation, the work would provide a general strategy for inferring effective many-body Hamiltonians from local spectroscopic data in systems with structural or environmental inhomogeneities. This is significant for quantum materials research, as it moves beyond the common assumption of uniform couplings and enables direct comparison between measured spectra and reconstructed models.

major comments (3)
  1. [Methods] Methods section: no description is provided of the machine learning training procedure, choice of validation sets, error bars on the reconstructed J(r), or controls for overfitting. This information is load-bearing for the claim that the models are inferred from the STS maps rather than being artifacts of the fitting process.
  2. [Results] Results section (spectra comparison): the assertion of quantitative reproduction of experimental spectra lacks reported metrics (e.g., RMS deviation, overlap integrals, or per-peak error) or explicit comparison tables/figures that would allow assessment of agreement quality across the claimed uniform and inhomogeneous cases.
  3. [Finite-size scaling] Finite-size scaling analysis: while the correct gap scaling with system length is claimed, the manuscript does not specify how the gap is extracted from the reconstructed Hamiltonian (e.g., exact diagonalization details or finite-size extrapolation procedure) or provide the raw gap values versus length for both experiment and model.
minor comments (2)
  1. [Abstract] The abstract states that the methodology is 'length-independent' but does not clarify whether this property was verified by training on chains of varying lengths and testing on unseen lengths.
  2. [Model definition] Notation for the spatially varying exchange J(r) should be defined explicitly in the main text, including how the local ML model parametrizes the spatial modulation.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the detailed and constructive report. The comments highlight important areas where additional methodological transparency and quantitative validation will strengthen the manuscript. We address each major comment below and will incorporate the requested information in a revised version.

read point-by-point responses

  1. Referee: [Methods] Methods section: no description is provided of the machine learning training procedure, choice of validation sets, error bars on the reconstructed J(r), or controls for overfitting. This information is load-bearing for the claim that the models are inferred from the STS maps rather than being artifacts of the fitting process.

    Authors: We agree that the current Methods section does not provide sufficient detail on these aspects. In the revised manuscript we will add a dedicated subsection describing the machine-learning training procedure (including network architecture, loss function, optimizer, and hyperparameters), the construction and use of validation sets, the method used to obtain error bars on the reconstructed J(r), and the regularization and cross-validation steps employed to control overfitting. These additions will directly support the claim that the inferred Hamiltonians are data-driven rather than fitting artifacts. revision: yes

  2. Referee: [Results] Results section (spectra comparison): the assertion of quantitative reproduction of experimental spectra lacks reported metrics (e.g., RMS deviation, overlap integrals, or per-peak error) or explicit comparison tables/figures that would allow assessment of agreement quality across the claimed uniform and inhomogeneous cases.

    Authors: We acknowledge that quantitative metrics and explicit comparison tables are absent. In the revision we will report RMS deviations between experimental and simulated spectra, peak-position errors, and (where appropriate) spectral overlap integrals. We will also add a supplementary table and/or figure that tabulates these metrics separately for the near-uniform and inhomogeneous chains, allowing direct assessment of agreement quality. revision: yes

  3. Referee: [Finite-size scaling] Finite-size scaling analysis: while the correct gap scaling with system length is claimed, the manuscript does not specify how the gap is extracted from the reconstructed Hamiltonian (e.g., exact diagonalization details or finite-size extrapolation procedure) or provide the raw gap values versus length for both experiment and model.

    Authors: We will expand the finite-size scaling section to specify the exact-diagonalization procedure (including Hilbert-space truncation and convergence criteria) used to obtain the excitation gap from each reconstructed Hamiltonian. We will also include the raw gap values versus chain length for both the experimental STS data and the model predictions, presented in a table or supplementary figure, together with the extrapolation procedure employed to confirm the scaling. revision: yes

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The abstract describes an ML-based reconstruction of spatially varying J(r) from inelastic STS maps, followed by verification that the resulting models reproduce the input spectra and recover gap scaling with length. No equations, self-citations, or derivation steps are supplied that reduce any claimed result to its inputs by construction. Reproduction of spectra is the expected outcome of a fitting procedure and does not render the reconstruction methodology circular; the central claim remains an independent inverse-problem technique validated against external experimental benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review supplies insufficient detail to enumerate free parameters, axioms, or invented entities; the approach implicitly assumes a Heisenberg model with spatially varying J_ij and that STS maps encode those couplings invertibly.

pith-pipeline@v0.9.1-grok · 5686 in / 1129 out tokens · 37324 ms · 2026-06-30T02:32:43.373765+00:00 · methodology

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

49 extracted references · 7 canonical work pages · 2 internal anchors

  1. [1]

    D. G. de Oteyza and T. Frederiksen, Carbon-based nanostructures as a versatile platform for tunableπ- magnetism, Journal of Physics: Condensed Matter34, 443001 (2022)

    2022

  2. [2]

    Zhang, B

    Y. Zhang, B. Fu, N. Li, J. Lu, and J. Cai, Advancements inπ-magnetism and precision engineering of carbon- based nanostructures, Chemistry – A European Journal 30, e202402765 (2024)

    2024

  3. [3]

    Mishra, G

    S. Mishra, G. Catarina, F. Wu, R. Ortiz, D. Jacob, K. Eimre, J. Ma, C. A. Pignedoli, X. Feng, P. Ruffieux, J. Fern´ andez-Rossier, and R. Fasel, Observation of frac- tional edge excitations in nanographene spin chains, Na- ture598, 287 (2021)

    2021

  4. [4]

    Chenxiao, G

    Z. Chenxiao, G. Catarina, J.-J. Zhang, J. Henriques, L. Yang, J. Ma, X. Feng, O. Gr¨ oning, P. Ruffieux, J. Fern´ andez, and R. Fasel, Tunable topological phases in nanographene-based spin-1/2 alternating-exchange heisenberg chains, Nature Nanotechnology19, 1789 (2024)

    2024

  5. [5]

    X. Su, Z. Ding, Y. Hong, N. Ke, K. Yan, C. Li, Y.-F. Jiang, and P. Yu, Fabrication of spin-1/2 heisenberg an- tiferromagnetic chains via combined on-surface synthesis and reduction for spinon detection, Nature Synthesis4, 694 (2025)

    2025

  6. [6]

    K. Sun, N. Cao, O. J. Silveira, A. O. Fumega, F. Hanin- dita, S. Ito, J. L. Lado, P. Liljeroth, A. S. Foster, and S. Kawai, On-surface synthesis of heisenberg spin-1/2 an- tiferromagnetic molecular chains, Science Advances11, eads1641 (2025)

    2025

  7. [7]

    X. Fu, L. Huang, K. Liu, J. C. G. Henriques, Y. Gao, X. Han, H. Chen, Y. Wang, C.-A. Palma, Z. Cheng, X. Lin, S. Du, J. Ma, J. Fern´ andez-Rossier, X. Feng, and H.-J. Gao, Building spin-1/2 antiferromagnetic heisen- berg chains with diaza-nanographenes, Nature Synthesis 4, 684 (2025)

    2025

  8. [8]

    Yuan, X.-Y

    Z. Yuan, X.-Y. Zhang, Y. Jiang, X. Qian, Y. Wang, Y. Liu, L. Liu, X. Liu, D. Guan, Y. Li, H. Zheng, C. Liu, J. Jia, M. Qin, P.-N. Liu, D.-Y. Li, and S. Wang, Fractional spinon quasiparticles in open-shell triangulene spin-1/2 chains, Journal of the American Chemical Soci- ety147, 5004 (2025)

    2025

  9. [9]

    C. Zhao, L. Yang, J. C. G. Henriques, M. Ferri-Cort´ es, G. Catarina, C. A. Pignedoli, J. Ma, X. Feng, P. Ruffieux, J. Fern´ andez-Rossier, and R. Fasel, Spin excitations in nanographene-based antiferromagnetic spin-1/2 heisen- berg chains, Nature Materials24, 722 (2025)

    2025

  10. [10]

    Vegliante, S

    A. Vegliante, S. Fern´ andez, R. Ortiz, M. Vilas-Varela, T. Y. Baum, N. Friedrich, F. Romero-Lara, A. Aguirre, K. Vaxevani, D. Wang, C. Garcia Fernandez, H. S. J. van der Zant, T. Frederiksen, D. Pe˜ na, and J. I. Pascual, Tuning the spin interaction in nonplanar organic dirad- icals through mechanical manipulation, ACS Nano18, 26514 (2024)

    2024

  11. [11]

    Catarina, E

    G. Catarina, E. Turco, N. Krane, M. Bommert, A. Ortega-Guerrero, O. Gr¨ oning, P. Ruffieux, R. Fasel, and C. A. Pignedoli, Conformational tuning of magnetic interactions in coupled nanographenes, Nano Letters24, 12536 (2024)

    2024

  12. [12]

    Krane, E

    N. Krane, E. Turco, A. Bernhardt, D. Jacob, G. Gan- dus, D. Passerone, M. Luisier, M. Jur´ ıˇ cek, R. Fasel, J. Fern´ andez-Rossier, and P. Ruffieux, Exchange inter- actions and intermolecular hybridization in a spin-1/2 nanographene dimer, Nano Letters23, 9353 (2023)

    2023

  13. [13]

    C. Zhao, Q. Huang, L. c. v. Valenta, K. Eimre, L. Yang, A. V. Yakutovich, W. Xu, J. Ma, X. Feng, M. Jur´ ıˇ cek, R. Fasel, P. Ruffieux, and C. A. Pignedoli, Tailoring mag- netism of graphene nanoflakes via tip-controlled dehydro- genation, Phys. Rev. Lett.132, 046201 (2024)

    2024

  14. [14]

    Carleo and M

    G. Carleo and M. Troyer, Solving the quantum many- body problem with artificial neural networks, Science 355, 602–606 (2017)

    2017

  15. [15]

    E. P. L. van Nieuwenburg, Y.-H. Liu, and S. D. Huber, Learning phase transitions by confusion, Nature Physics 13, 435–439 (2017)

    2017

  16. [16]

    Carrasquilla and R

    J. Carrasquilla and R. G. Melko, Machine learning phases of matter, Nature Physics13, 431–434 (2017)

    2017

  17. [17]

    Durrer, B

    R. Durrer, B. Kratochwil, J. Koski, A. Landig, C. Re- ichl, W. Wegscheider, T. Ihn, and E. Greplova, Au- tomated tuning of double quantum dots into specific charge states using neural networks, Phys. Rev. Appl. 13, 054019 (2020)

    2020

  18. [19]

    I.-J. Chen, M. Aapro, A. Kipnis, A. Ilin, P. Liljeroth, and A. S. Foster, Precise atom manipulation through deep reinforcement learning, Nature Communications 13, 10.1038/s41467-022-35149-w (2022)

    work page doi:10.1038/s41467-022-35149-w 2022

  19. [20]

    N. Wu, M. Aapro, J. S. Jestil¨ a, R. Drost, M. M. Garc´ ıa, T. Torres, F. Xiang, N. Cao, Z. He, G. Bottari, P. Lil- jeroth, and A. S. Foster, Precise large-scale chemical transformations on surfaces: Deep learning meets scan- ning probe microscopy with interpretability, Journal of the American Chemical Society147, 1240–1250 (2024)

    2024

  20. [22]

    Lupi and J

    G. Lupi and J. L. Lado, Hamiltonian-learning quantum magnets with nonlocal impurity tomography, Phys. Rev. Appl.23, 054077 (2025)

    2025

  21. [23]

    Karjalainen, Z

    N. Karjalainen, Z. Lippo, G. Chen, R. Koch, A. O. Fumega, and J. L. Lado, Hamiltonian inference from dy- namical excitations in confined quantum magnets, Phys. Rev. Appl.20, 024054 (2023)

    2023

  22. [24]

    Karjalainen, G

    N. Karjalainen, G. Lupi, R. Koch, A. O. Fumega, and J. L. Lado, Hamiltonian learning quantum magnets with 6 dynamical impurity tomography, arXiv:2510.18613

    work page arXiv

  23. [25]

    Cross-Platform Autonomous Control of Minimal Kitaev Chains

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

    work page internal anchor Pith review Pith/arXiv arXiv

  24. [26]

    G. Lupi, A. O. Fumega, M. Amini, R. Drost, P. Lil- jeroth, and J. L. Lado, Molecular hamiltonian learn- ing from setpoint-dependent scanning tunneling spec- troscopy (2026), arXiv:2601.19371 [cond-mat.mes-hall]

    work page arXiv 2026

  25. [27]

    Hernandes, J

    V. Hernandes, J. Rogers, R. Koch, T. Spriggs, B. Und- seth, A. Chatterjee, L. M. K. Vandersypen, and E. Gre- plova, Reconstructing quantum dot charge stability dia- grams with diffusion models, arXiv:2603.26432

    work page arXiv

  26. [28]

    Hamiltonian learning for spin-spiral moir\'e magnets from electronic magnetotransport

    F. Nigmatulin, G. Lupi, J. L. Lado, and Z. Sun, Hamil- tonian learning for spin-spiral moir´ e magnets from elec- tronic magnetotransport (2026), arXiv:2604.02959 [cond- mat.mes-hall]

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  27. [29]

    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)

    2017

  28. [30]

    Koch and J

    R. Koch and J. L. Lado, Designing quantum many-body matter with conditional generative adversarial networks, Phys. Rev. Res.4, 033223 (2022)

    2022

  29. [31]

    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)

    2023

  30. [32]

    Valenti, G

    A. Valenti, G. Jin, J. L´ eonard, S. D. Huber, and E. Gre- plova, Scalable hamiltonian learning for large-scale out- of-equilibrium quantum dynamics, Phys. Rev. A105, 023302 (2022)

    2022

  31. [33]

    L. Che, C. Wei, Y. Huang, D. Zhao, S. Xue, X. Nie, J. Li, D. Lu, and T. Xin, Learning quantum hamiltonians from single-qubit measurements, Phys. Rev. Res.3, 023246 (2021)

    2021

  32. [34]

    Simard, A

    O. Simard, A. Dawid, J. Tindall, M. Ferrero, A. M. Sen- gupta, and A. Georges, Learning interactions between Rydberg atoms, PRX Quantum6, 030324 (2025)

    2025

  33. [35]

    Bairey, I

    E. Bairey, I. Arad, and N. H. Lindner, Learning a local Hamiltonian from local measurements, Phys. Rev. Lett. 122, 020504 (2019)

    2019

  34. [36]

    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, Nat. Rev. Phys.5, 141 (2023)

    2023

  35. [37]

    J. A. Sobral, S. Obernauer, S. Turkel, A. N. Pasupathy, and M. S. Scheurer, Machine learning the microscopic form of nematic order in twisted double-bilayer graphene, Nature Communications14, 10.1038/s41467-023-40684- 1 (2023)

    work page doi:10.1038/s41467-023-40684- 2023

  36. [38]

    R. Koch, R. Drost, P. Liljeroth, and J. L. Lado, Hamil- tonian learning of triplon excitations in an artificial nanoscale molecular quantum magnet, Nano Letters25, 13435 (2025)

    2025

  37. [39]

    Heinrich, J

    A. Heinrich, J. Gupta, C. Lutz, and D. Eigler, Single- atom spin-flip spectroscopy, Science (New York, N.Y.) 306, 466 (2004)

    2004

  38. [40]

    C. F. Hirjibehedin, C. P. Lutz, and A. J. Heinrich, Spin coupling in engineered atomic structures, Science312, 1021 (2006)

    2006

  39. [41]

    Ternes, Spin excitations and correlations in scan- ning tunneling spectroscopy, New Journal of Physics17, 063016 (2015)

    M. Ternes, Spin excitations and correlations in scan- ning tunneling spectroscopy, New Journal of Physics17, 063016 (2015)

    2015

  40. [42]

    The details of dataset generation and preprocessing are provided in the Supplemental Information (SI)

  41. [43]

    Systematic benchmarks of the trained algorithm in the SI

  42. [44]

    Further details on the synthesis are provided in the SI

  43. [45]

    Further experimental details are provided in the SI

  44. [46]

    See Supplementary Information for a detailed compari- son, including reconstructed maps and quantitative error analysis

  45. [47]

    Lupi, Code repository for ”learning inhomo- geneous heisenberg hamiltonians in nanographene spin chains”,https://github.com/GretaLupi/ Inhomogeneous-Heisenberg-HL(2026)

    G. Lupi, Code repository for ”learning inhomo- geneous heisenberg hamiltonians in nanographene spin chains”,https://github.com/GretaLupi/ Inhomogeneous-Heisenberg-HL(2026)

    2026

  46. [48]

    Holzner, A

    A. Holzner, A. Weichselbaum, I. P. McCulloch, U. Schollw¨ ock, and J. von Delft, Chebyshev matrix prod- uct state approach for spectral functions, Phys. Rev. B 83, 195115 (2011)

    2011

  47. [49]

    J. L. Lado and O. Zilberberg, Topological spin excita- tions in harper-heisenberg spin chains, Phys. Rev. Res. 1, 033009 (2019)

    2019

  48. [50]

    Fishman, S

    M. Fishman, S. R. White, and E. M. Stoudenmire, The ITensor Software Library for Tensor Network Calcula- tions, SciPost Phys. Codebases , 4 (2022)

    2022

  49. [51]

    J. L. Lado, https://github.com/joselado/dmrgpy (2025). 7 Supplemental Material SI. SYNTHESIS AND SPECTROSCOPY DETAILS This section provides the synthesis and measurement details for Chain H, while the synthesis and spectroscopy of Chain IH are reported in Ref. [9]. FIG. 5. On-surface synthesis route used to fabricate the Olympicene-based Heisenberg chains...

    2025