Pith. sign in

REVIEW 3 major objections 4 minor 56 references

AIMS: An uncertainty-aware AI experimentalist for quantum matter

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

Pith's one-line read An AI agent running a real cryogenic microscope claims that the surprisingly stable half-filled electron crystal in twisted bilayer MoSe2 melts slowly because electron hopping renormalizes its energy, not because of stronger classical order

desk verdict A genuinely integrated closed-loop AI experimentalist on a real cryogenic instrument, with a credible navigation and site-selection story — but the quantum-mechanism conclusion rides on dielectric parameters that are optimized differently across methods, so it needs a sensitivity analysis before it should be read as a prediction. read the letter →

arxiv 2607.16544 v1 pith:VHAR4PWP submitted 2026-07-17 cond-mat.str-el cond-mat.mes-hall

classification cond-mat.str-elcond-mat.mes-hall PACS 07.79.-v71.27.+a
keywords AIexperimentalistuncertainty-awareagentmicrowaveimpedancemicroscopygeneralizedWignercrystaltwistedbilayerMoSe2quantumfluctuationsmeltingtemperatureclosed-loopdiscovery
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper introduces AIMS (AI agent for Inference and Measurement in Science), a closed-loop AI experimentalist that operates a cryogenic microwave impedance microscope and converts three kinds of uncertainty—where the tip is, where to measure, what physics explains the data—into concrete next actions. In the physical case study, twisted bilayer MoSe2 shows an unusual thermal hierarchy: the half-filled generalized Wigner crystal (an electron crystal at one electron per two moiré sites) melts near 30 K, while the one-third- and two-thirds-filled crystals melt near 17 K. The paper claims this is not explained by classical charge order: a hopping-free Monte Carlo model inverts the hierarchy, while models with electron hopping reproduce it. The conclusion is that quantum fluctuations—electron hopping—are what make the half-filled stripe exceptionally robust. The wider point is that an AI experimentalist can work in a messy real laboratory, recovering from failed position estimates, choosing optimal measurement sites, and updating mechanism rankings as new evidence arrives.

What carries the argument

The load-bearing physical mechanism is the electron-hopping term t in the moiré-lattice model of twisted bilayer MoSe2: with t=0, classical Monte Carlo predicts ν=1/2 as the least thermally stable fractional crystal, opposite to experiment; turning on hopping reverses the ranking, and exact-diagonalization/finite-temperature-Lanczos calculations show hopping raises the ν=1/2 melting temperature while suppressing ν=1/3 and ν=2/3. Around this sits AIMS itself—a three-loop agent that converts position uncertainty, sample inhomogeneity, and interpretational ambiguity into targeted measurements, using microwave-impedance dip depth as the experimental melting observable.

What would settle it

Extend the temperature series on the same twisted MoSe2 device above 20 K and measure whether the ν=1/2 dip actually melts near 30 K; if it melts at or below roughly 17 K, the claimed hierarchy is false. Alternatively, run the classical Monte Carlo and Hartree-Fock models with dielectric constants measured independently on the actual device (not fitted): if the classical model then reproduces the observed hierarchy, the attribution to electron hopping is falsified.

Watch

Extended reading notes

Core claim

On its own terms, the central discovery is that the melting hierarchy Tm(ν=1/2) > Tm(ν=1/3) ≃ Tm(ν=2/3) in twisted bilayer MoSe2 is not inherited from stronger classical charge order but enabled by electron hopping. The paper shows that classical Monte Carlo with zero hopping predicts ν=1/2 as the least stable state, opposite to experiment, while Hartree-Fock and exact-diagonalization/finite-temperature-Lanczos calculations with finite hopping reproduce the observed ordering, with hopping raising the ν=1/2 melting scale from roughly 5 K to roughly 30 K while lowering the ν=1/3 and ν=2/3 scales. Alongside this, AIMS demonstrates closed-loop experimental agency: it relocates samples after cryo

Load-bearing premise

The hierarchy attribution depends on treating MIM dip depth as a monotone proxy for thermodynamic melting temperature, and on dielectric parameters in the model calculations that are in part chosen rather than independently measured—if those give way, the quantum-fluctuation ranking loses its footing.

Editorial extensions

If this is right

  • If the quantum-fluctuation account is right, the ν=1/2 generalized Wigner crystal in twisted MoSe2 should remain the most thermally robust of the three fractional states up to at least about 30 K, with a melting curve following the same power-law form as the neighboring states.
  • The navigation results imply that a similar agent could reduce sample-locating time after cooldown from about ten hours to about four hours on other cryogenic scanning probes equipped with directional marker patterns and recovery tools.
  • The measurement-loop results imply that optimal spectroscopy locations in inhomogeneous moiré devices can be found automatically by combining twist-angle maps with a quantitative correlated-feature score, with independent human choice falling within about 200 nm of the agent's selection.
  • The discovery loop establishes a template for mechanism attribution under ambiguous physics: instead of a binary classical/quantum verdict, the agent ranks hypotheses, identifies the specific missing evidence, and updates posterior probabilities with new calculations and measurements.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If hopping-induced renormalization is a general feature of moiré electron crystals, other fractional fillings with stripe-like order might show similar enhancement; the paper tests only ν=1/3, 1/2, and 2/3 in this material.
  • The same three-loop architecture—position recovery, targeted measurement, evidence-driven attribution—could transfer to other sparse-signal, drift-prone instruments beyond microwave impedance microscopy, since none of the loop designs depend on the specific probe physics.
  • The paper's use of a censored 30.5 K melting temperature suggests a direct falsifier: a higher-temperature-capable measurement of the ν=1/2 dip would confirm or refute the extrapolation without relying on the shared power law.
  • One could test the mechanism ranking without fitted parameters by independently measuring the dielectric constants of the hBN-encapsulated device and repeating the classical and quantum calculations; the paper does not report such a parameter-free test.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 4 minor

Summary. This paper introduces AIMS, an LLM-based closed-loop agent for cryogenic microwave impedance microscopy, and demonstrates it on three nested tasks: locating a sample after cryogenic displacement, choosing the best spectroscopy site in a disordered twisted bilayer MoSe2 device, and attributing the anomalous melting hierarchy Tm(ν=1/2)>Tm(ν=1/3)≈Tm(ν=2/3) to quantum-fluctuation-renormalized melting rather than classical charge order. The agent uses particle-filter localization with uncertainty-triggered GP-regression recovery, a GWC score for site selection, and a hypothesis-ranking loop that invokes classical Monte Carlo, Hartree-Fock, exact-diagonalization/finite-temperature Lanczos calculations, and registered AFM as evidence. The paper claims significant time savings in navigation, agreement of the selected site with an independent human grid, and a physical mechanism in which electron hopping stabilizes the half-filled stripe.

Significance. The paper advances the benchmark for AI experimentalists in quantum materials by making uncertainty actionable rather than merely automating scans. Its strengths include explicit failure detection and recovery in navigation, independent human benchmarks for site selection, a registered structural control that excludes twist-domain morphology as the primary explanation, and blind evaluation against human reports. If the mechanism claim survives parameter-sensitivity testing, the work would provide a compelling demonstration that an LLM-driven agent can close the loop from instrument control to physical interpretation. The main weight of the paper falls on the discovery-loop conclusion, so the adequacy of the model comparison rather than the autonomy demonstration determines its contribution.

major comments (3)
  1. [Fig. 4(d)-(e) / 'Resolving GWC melting mechanisms'] The HF model uses Bayesian-optimized εr=14.2, ε⊥=6.0, while the ED/FTLM calculation that produces the ~30 K ν=1/2 melting scale uses εr=3, ε⊥=6.0. Since εr sets the Coulomb scale V, the ED/FTLM curve is not sampling the same physical regime as the HF optimum, and no sensitivity analysis is shown for the hierarchy across the plausible εr range. The quantitative agreement between ED/FTLM and the extrapolated 30.5±6 K is therefore potentially a product of parameter selection. Please provide a scan over (εr, ε⊥) for the ED/FTLM and HF models, or justify the different values from independent measurements, before assigning the hierarchy to quantum fluctuations.
  2. [Methods, 'Melting temperature extraction'] Tm(ν=1/2)=30.5±6.0 K is an extrapolation: at 20 K the dip retains ~48% of its maximum depth, and the power law O(T)=A(1−T/Tc)^β with β≈0.7 is fixed by the in-range states rather than independently measured. Moreover, dip-depth loss is assumed to be a monotone proxy for thermodynamic melting without validation. The match of ED/FTLM to 30 K thus partly reflects the assumed extrapolation. I recommend reporting the raw depth-vs-T data, an uncertainty budget for β, and, if possible, an independent melting probe for the ν=1/2 state.
  3. [Fig. 4(d): Bayesian-optimizing HF parameters] The finite-hopping Hartree-Fock model is said to reproduce the hierarchy 'under identical priors and likelihoods' but with Bayesian-optimized parameters. Fitting parameters to the same melting hierarchy and then ranking mechanisms with that model is circular unless there is an out-of-sample prediction or a prior-predictive check. Please state which observables constrained εr and ε⊥, and show that the ranking is stable when parameters are varied within their posterior/prior range.
minor comments (4)
  1. [Eq. (2)] The GWC score weights w1..w4 are set by hand. A sensitivity analysis would strengthen the measurement-selection claim, although the independent human grid agreement mitigates this concern.
  2. [Fig. 2(g) and navigation results] The reported ~4 h vs ~10 h navigation-time comparison appears to be a single realization. Repeated trials are mentioned in the SI; state the number of runs and variability in the main text.
  3. [Data and Code Availability] The statement that all data and code are 'available from the corresponding authors upon request' is insufficiently strong for an AI-agent paper. Deposit prompts, MCP tools, analysis scripts, and raw/processed data in a public repository to allow independent re-analysis.
  4. [Fig. 4(h)] Blind scoring by three physicists needs a more explicit rubric and inter-rater agreement information; three raters is a small sample, so the comparison should be framed accordingly.

Circularity Check

3 steps flagged · score 6.0 of 10

Discovery-loop validation is partly circular: model Tm values are explicitly 'optimized' to the experimental Tm, and the two quantum models use inconsistent fitted dielectric constants, so the quantum-fluctuation attribution is not a parameter-free prediction.

  1. fitted input called prediction [Figure 4(g) caption; discovery-loop second inference pass]
    "The inset shows the optimized T_m from MC, HF, ED/FTLM models compared to the experimental T_m for each filling."

    The mechanism ranking is justified by agreement between model and experiment, but the figure explicitly states that the model Tm values were 'optimized' to the experimental Tm. A fit to the benchmark cannot then serve as independent evidence for the benchmark. Since the quantum-fluctuation mechanism is ranked 'with all evidence from pass 1 and 2' and this inset is the quantitative closure, the preference for the quantum hypothesis is partly forced by the fit rather than by prediction.

  2. fitted input called prediction [Figure 4(d)-(e) captions; 'Resolving generalized Wigner crystal melting mechanisms' section]
    "simulated based on the Hartree-Fock model with Bayesian optimized parameters ε_r = 14.2 and ε⊥ = 6.0 ... obtained from exact diagonalization and finite-temperature Lanczos method at ε_r = 3 and ε⊥ = 6.0."

    The HF reproduction of the hierarchy uses Bayesian-optimized dielectric parameters, while the ED/FTLM curve that produces the key 'approximately 30 K at the physical hopping' uses a different ε_r (3 vs 14.2). No sensitivity analysis is shown across this range, so the match to the extrapolated experimental Tm(ν=1/2)=30.5±6.0 K is parameter-selected, not a robust first-principles prediction. The ED/FTLM point thereby inherits the optimization of the previous step.

1 more flagged steps
  1. other [Methods: Melting temperature extraction]
    "For ν=1/2, which retains ∼48% of its maximum depth at 20 K, T_m is obtained by extrapolation with a shared power law O(T)=A(1−T/T_c)^β, with β≈0.7 fixed by the in-range states."

    The 'experimental' validation target for the half-filled state is not directly measured but extrapolated using a power-law form whose exponent is fixed by the other, in-range states. Comparing the ED/FTLM 'physical hopping' point to this extrapolated number is therefore comparing two fitted quantities; the 30 K agreement is not an independent observation of a melting transition.

full rationale

Navigation and measurement loops are empirically controlled (particle-filter localization cross-checks, human SSIM/denser-grid benchmarks, blind human report scoring) and show no circularity. The discovery loop's qualitative structure is also not circular: the classical t=0 Monte Carlo result (ν=1/2 least robust) and the ED/FTLM trend (hopping strongly raises Tm(1/2) while suppressing neighbors) are model outputs with independent content, and the registered AFM measurement is a genuine morphological control. The circularity is confined to the quantitative validation layer. Figure 4(g) explicitly presents 'optimized Tm' from MC, HF, and ED/FTLM alongside experimental Tm as closure evidence; optimizing the models to the experimental melting temperatures and then using the resulting agreement to rank the quantum mechanism is a fitted-input-called-prediction step. This is compounded by the inconsistent dielectric constants (ε_r=14.2 in HF vs ε_r=3 in ED/FTLM) with no sensitivity analysis, and by the fact that Tm(ν=1/2) itself is an extrapolated, power-law-derived quantity rather than a measured transition. These issues weaken the quantitative claim that hopping 'raises the melting scale ... to approximately 30 K at the physical hopping,' but they do not eliminate the qualitative mechanism ranking, nor do they affect the navigation/measurement loop results. Overall score 6: partial circularity in the central quantitative discovery claim, while substantial independent content remains.

Assumptions & free parameters 5 free parameters · 5 assumptions · 0 invented entities

The central systems claim rests mainly on experimental demonstration and benchmark comparisons; the physical mechanism claim rests on fitted model parameters and analysis thresholds. The most consequential free parameters are the dielectric constants used in the HF and ED/FTLM calculations, which are not independently measured and differ between the two methods. The melting-temperature extraction adds a power-law exponent fitted to the data. No new physical entities are introduced.

free parameters (5)
  • HF dielectric parameters εr, ε⊥ = εr=14.2, ε⊥=6.0
    Bayesian-optimized in the Hartree-Fock model used to rank melting mechanisms (Fig. 4d). Not independently measured; potentially optimized to match the experimental hierarchy.
  • ED/FTLM dielectric parameters εr, ε⊥ = εr=3, ε⊥=6.0
    Chosen for the exact-diagonalization/finite-temperature Lanczos calculations (Fig. 4e); inconsistent with the HF values, and the choice affects the melting-scale magnitudes.
  • Power-law exponent β = 0.7
    Used to extrapolate Tm(ν=1/2)=30.5±6.0 K; fixed by fitting the in-range ν=1/3 and 2/3 states, making the extrapolated 1/2 value partly a fitted quantity.
  • GWC score weights w1,w2,w3,w4 = 1, 0.3, 1, 1
    Hand-picked weights in Eq. (2) for site scoring; they directly determine which measurement position AIMS selects, and sensitivity to these weights is not reported.
  • Melting extraction thresholds = 10% depth criterion, 28% shoulder window, 2× noise threshold
    Analysis thresholds used to define dip significance and Tm; the extracted melting hierarchy may depend on these choices, and no sensitivity analysis is given.
assumptions (5)
  • domain assumption MIM dip depth is a monotone proxy for thermodynamic melting temperature of GWC states
    The entire melting hierarchy and the Tm values are derived from normalized dip depth (Methods: Melting temperature extraction). The paper explicitly calls dip depth an 'imperfect proxy,' so the central physical ranking rests on this unproven monotonicity.
  • domain assumption Classical Monte Carlo, Hartree-Fock, and ED/FTLM models capture the relevant physics of GWC melting in twisted MoSe2
    Discovery-loop mechanism ranking treats these calculations as faithful representations of the real system; no direct validation of the models against the full experimental data is provided beyond the selected Tm comparisons.
  • domain assumption The LLM (Claude Opus 4.5/4.8) provides reliable tool execution, self-assessment of uncertainty, and calibrated Bayesian posteriors
    The AIMS workflow depends on the LLM correctly flagging uncertain positions, choosing recovery strategies, and producing 'calibrated' posterior probabilities. No formal guarantees or detailed reliability statistics for the LLM reasoning are provided.
  • standard math Twist angle extraction formula Eq. (1) holds in the small-angle, single-gated limit
    The twist-angle disorder map and site-selection scoring rely on Eq. (1), which is stated to be valid for single-gated hexagonal twisted heterostructures in the small-angle limit; deviations could bias the twist-angle map.
  • standard math Particle filter/template matching and Gaussian-process reconstruction correctly localize the tip
    Navigation-loop success depends on these algorithmic localization methods. The paper benchmarks them against CNN and human navigation, providing partial support, but no formal correctness guarantee.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AIMS: An uncertainty-aware AI experimentalist for quantum matter." pith.science (2026). https://pith.science/paper/VHAR4PWP

@misc{pith2026260716544,
  author       = {Pith},
  title        = {Pith review of: AIMS: An uncertainty-aware AI experimentalist for quantum matter},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VHAR4PWP}},
  note         = {Machine review of arXiv:2607.16544}
}
abstract

Autonomous scientific agents are beginning to accelerate discovery, but most demonstrations operate in digital or highly structured settings where the objects, actions, and objectives are largely predefined. Quantum materials experiments pose a harder problem where uncertainties involve: the instrument state can drift, the useful signal may occupy only rare regions of an inhomogeneous sample, and the physical mechanism is often under-determined. Here we introduce the AI agent for Inference and Measurement in Science (AIMS), an uncertainty-aware closed-loop AI experimentalist for cryogenic microwave impedance microscopy that converts uncertainty into experimental action. AIMS links three nested loops: navigation under uncertain perception, measurement selection under sample inhomogeneity, and scale-resolved mechanism attribution under ambiguous physics. In navigation, it relocates the sample after cryogenic displacement, flags unreliable position estimates, and invokes recovery strategies, significantly reducing sample-locating time. In measurement, it maps twist angle distribution and generalized Wigner crystal score of twisted bilayer MoSe$_2$ to identify regions with the strongest correlated response. In discovery, AIMS asks not whether melting is simply classical or quantum, but how the competition of Coulomb repulsion, hopping, and other energy scales shapes the observed hierarchy. By testing the classical limit, varying hopping and Coulomb scales, and preserving sample morphology as a secondary testable variable, AIMS prioritizes a quantum-fluctuation-renormalized origin of the anomalously robust $\nu = 1/2$ crystal. AIMS demonstrates uncertainty-aware experimental agency for quantum matter with perception recovery, measurement choice, and energy-scale-resolved mechanism attribution in one closed loop.

Figures

Figures reproduced from arXiv: 2607.16544 by the authors.

Figure 1
Figure 1. FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4 [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

56 extracted references · 4 canonical work pages

  1. [1]

    Wang, H.et al.Scientific discovery in the age of artificial intelligence.Nature620, 47–60 (2023)

  2. [2]

    A., MacKnight, R., Kline, B

    Boiko, D. A., MacKnight, R., Kline, B. & Gomes, G. Au- tonomous chemical research with large language models. Nature624, 570–578 (2023)

  3. [3]

    M.et al.Augmenting large language mod- els with chemistry tools.Nat

    Bran, A. M.et al.Augmenting large language mod- els with chemistry tools.Nat. Mach. Intell.6, 525–535 (2024)

  4. [4]

    Dai, T.et al.Autonomous mobile robots for exploratory synthetic chemistry.Nature635, 890–897 (2024)

  5. [5]

    C., Collison, C

    Ramos, M. C., Collison, C. J. & White, A. D. A re- view of large language models and autonomous agents in chemistry.Chem. Sci.16, 2514–2572 (2025)

  6. [6]

    Preprint athttps://doi.org/10.48550/arXiv

    Ni, Z.et al.MatPilot: an LLM-enabled AI materials scientist under the framework of human-machine collab- oration. Preprint athttps://doi.org/10.48550/arXiv. 2411.08063(2024)

  7. [7]

    P.et al.Self-driving laboratory for ac- celerated discovery of thin-film materials.Sci

    MacLeod, B. P.et al.Self-driving laboratory for ac- celerated discovery of thin-film materials.Sci. Adv.6, eaaz8867 (2020)

  8. [8]

    J.et al.An autonomous laboratory for the accelerated synthesis of inorganic materials.Nature 624, 86–91 (2023)

    Szymanski, N. J.et al.An autonomous laboratory for the accelerated synthesis of inorganic materials.Nature 624, 86–91 (2023)

Show all 56 references
  1. [9]

    Rev.124, 9633–9732 (2024)

    Tom, G.et al.Self-driving laboratories for chemistry and materials science.Chem. Rev.124, 9633–9732 (2024)

  2. [10]

    Zhang, Z.et al.A multimodal robotic platform for multi- element electrocatalyst discovery.Nature647, 390–396 (2025)

  3. [11]

    Neural Inf

    Wei, J.et al.Chain-of-thought prompting elicits reason- ing in large language models.Adv. Neural Inf. Process. Syst.35, 24824–24837 (2022)

  4. [12]

    Yao, S.et al.ReAct: synergizing reasoning and acting in language models. InProc. 11th International Conference on Learning Representations (ICLR)(2023)

  5. [13]

    Preprint athttps: //doi.org/10.48550/arXiv.2605.18407(2026)

    Shi, L.et al.Qumus: realization of an embodied AI quantum material experimentalist. Preprint athttps: //doi.org/10.48550/arXiv.2605.18407(2026)

  6. [14]

    Preprint athttps://doi

    Li, X.et al.An AI-driven robotic system for two- dimensional hetero-assemblies. Preprint athttps://doi. org/10.48550/arXiv.2605.20420(2026)

  7. [15]

    ACS Nano19, 35493–35502 (2025)

    Yang, J.et al.Zero-shot autonomous microscopy for scalable and intelligent characterization of 2D materials. ACS Nano19, 35493–35502 (2025)

  8. [16]

    Qu, Y.et al.CRISPR-GPT for agentic automation of gene-editing experiments.Nat. Biomed. Eng.10, 245– 258 (2026)

  9. [17]

    H., Zhou, T., Chan, H

    Vriza, A., Prince, M. H., Zhou, T., Chan, H. & Cherukara, M. J. Operating advanced scientific instru- ments with AI agents that learn on the job.npj Comput. Mater.12, 160 (2026)

  10. [18]

    C., ˇSulc, A

    Hellert, T., Bertwistle, D., Leemann, S. C., ˇSulc, A. & Venturini, M. Agentic artificial intelligence for multistage physics experiments at a large-scale user facility particle accelerator.Phys. Rev. Res.8, L012017 (2026)

  11. [19]

    Chen, Z.et al.An agentic artificially intelligent X-ray scientist (2025)

  12. [20]

    Uri, A.et al.Mapping the twist-angle disorder and Lan- dau levels in magic-angle graphene.Nature581, 47–52 (2020)

  13. [21]

    P.et al.Strain fields in twisted bilayer graphene.Nat

    Kazmierczak, N. P.et al.Strain fields in twisted bilayer graphene.Nat. Mater.20, 956–963 (2021)

  14. [22]

    Mater.23, 1070–1076 (2024)

    Li, Q.et al.Evolution of the flat band and the role of lattice relaxations in twisted bilayer graphene.Nat. Mater.23, 1070–1076 (2024)

  15. [23]

    C.et al.Imaging supermoir´ e relaxation in helical trilayer graphene.Nat

    Hoke, J. C.et al.Imaging supermoir´ e relaxation in helical trilayer graphene.Nat. Mater.25, 775–781 (2026)

  16. [24]

    Nanotechnol.15, 750–754 9 (2020)

    Sung, J.et al.Broken mirror symmetry in exci- tonic response of reconstructed domains in twisted MoSe2/MoSe2 bilayers.Nat. Nanotechnol.15, 750–754 9 (2020)

  17. [25]

    Phys.17, 720– 725 (2021)

    Shabani, S.et al.Deep moir´ e potentials in twisted tran- sition metal dichalcogenide bilayers.Nat. Phys.17, 720– 725 (2021)

  18. [26]

    Wigner crystals in flat band 2D electron systems.Nat

    Shayegan, M. Wigner crystals in flat band 2D electron systems.Nat. Rev. Phys.4, 212–213 (2022)

  19. [27]

    & Kim, E.-A

    Matty, M. & Kim, E.-A. Melting of generalized Wigner crystals in transition metal dichalcogenide heterobilayer moir´ e systems.Nat. Commun.13, 7098 (2022)

  20. [28]

    Zhou, Y., Sheng, D. N. & Kim, E.-A. Quantum melting of generalized Wigner crystals in transition metal dichalco- genide moir´ e systems.Phys. Rev. Lett.133, 156501 (2024)

  21. [29]

    Nuckolls, K. P. & Yazdani, A. A microscopic perspective on moir´ e materials.Nat. Rev. Mater.9, 460–480 (2024)

  22. [30]

    & Changlani, H

    Kumar, A., Sherif, S., Elser, V. & Changlani, H. J. Melt- ing temperature shifts from quantum fluctuations in gen- eralized Wigner crystals. Preprint athttps://doi.org/ 10.48550/arXiv.2604.19870(2026)

  23. [31]

    Lai, K.et al.Mesoscopic percolating resistance network in a strained manganite thin film.Science329, 190–193 (2010)

  24. [32]

    Y.et al.Mobile metallic domain walls in an all-in-all-out magnetic insulator.Science350, 538–541 (2015)

    Ma, E. Y.et al.Mobile metallic domain walls in an all-in-all-out magnetic insulator.Science350, 538–541 (2015)

  25. [33]

    Cui, Y.-T.et al.Unconventional correlation between quantum Hall transport quantization and bulk state fill- ing in gated graphene devices.Phys. Rev. Lett.117, 186601 (2016)

  26. [34]

    Natl Acad

    Allen, M.et al.Visualization of an axion insulating state at the transition between 2 chiral quantum anomalous Hall states.Proc. Natl Acad. Sci. USA116, 14511–14515 (2019)

  27. [35]

    E., Ma, E

    Barber, M. E., Ma, E. Y. & Shen, Z.-X. Microwave impedance microscopy and its application to quantum materials.Nat. Rev. Phys.4, 61–74 (2021)

  28. [36]

    Adv.5, eaat8799 (2019)

    Shi, Y.et al.Imaging quantum spin Hall edges in mono- layer wte2.Sci. Adv.5, eaat8799 (2019)

  29. [37]

    Ji, Z.et al.Local probe of bulk and edge states in a fractional Chern insulator.Nature635, 578–583 (2024)

  30. [38]

    Model context proto- col specification.https://modelcontextprotocol.io/ specification/2025-11-25(2025)

    Model Context Protocol. Model context proto- col specification.https://modelcontextprotocol.io/ specification/2025-11-25(2025). Version 2025-11-25

  31. [39]

    S., Maskell, S., Gordon, N

    Arulampalam, M. S., Maskell, S., Gordon, N. & Clapp, T. A tutorial on particle filters for online nonlinear/non- Gaussian Bayesian tracking.IEEE Trans. Signal Process. 50, 174–188 (2002)

  32. [40]

    Lewis, J. P. Fast normalized cross-correlation. InProc. Vision Interface, 120–123 (1995)

  33. [41]

    & Triggs, B

    Dalal, N. & Triggs, B. Histograms of oriented gradients for human detection. InProc. IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), vol. 1, 886–893 (IEEE, 2005)

  34. [42]

    & Hinton, G

    Krizhevsky, A., Sutskever, I. & Hinton, G. E. ImageNet classification with deep convolutional neural networks. In Adv. Neural Inf. Process. Syst., vol. 25 (2012)

  35. [43]

    P.et al.Fast scanning probe microscopy via machine learning: non-rectangular scans with com- pressed sensing and Gaussian process optimization

    Kelley, K. P.et al.Fast scanning probe microscopy via machine learning: non-rectangular scans with com- pressed sensing and Gaussian process optimization. Small16, 2002878 (2020)

  36. [44]

    Mater.6, 21 (2020)

    Ziatdinov, M.et al.Imaging mechanism for hyperspectral scanning probe microscopy via Gaussian process mod- elling.npj Comput. Mater.6, 21 (2020)

  37. [45]

    Fu, M.et al.Accelerated nano-optical imaging through sparse sampling.Nano Lett.24, 2149–2156 (2024)

  38. [46]

    Commun.16, 4058 (2025)

    Zong, Q.-J.et al.Quantum melting of generalized elec- tron crystal in twisted bilayer MoSe2.Nat. Commun.16, 4058 (2025)

  39. [47]

    C.et al.Mott and generalized Wigner crystal states in WSe2/WS2 moir´ e superlattices.Nature579, 359–363 (2020)

    Regan, E. C.et al.Mott and generalized Wigner crystal states in WSe2/WS2 moir´ e superlattices.Nature579, 359–363 (2020)

  40. [48]

    Xu, Y.et al.Correlated insulating states at fractional fill- ings of moir´ e superlattices.Nature587, 214–218 (2020)

  41. [49]

    Phys.17, 715–719 (2021)

    Huang, X.et al.Correlated insulating states at fractional fillings of the WS 2/WSe2 moir´ e lattice.Nat. Phys.17, 715–719 (2021)

  42. [50]

    Preprint at https://doi.org/10.48550/arXiv.2504.17970(2025)

    Yang, D.et al.Correlated insulating states in slow Dirac fermions on a honeycomb moir´ e superlattice. Preprint at https://doi.org/10.48550/arXiv.2504.17970(2025)

  43. [51]

    Huang, X.et al.Measurements of correlated insulator gaps in a transition-metal dichalcogenide moir´ e superlat- tice.Nano Lett.25, 9214–9220 (2025)

  44. [52]

    Preprint athttps://doi.org/10.48550/ arXiv.2311.04211(2023)

    Ji, Z.et al.Harnessing excitons at the nanoscale– photoelectrical platform for quantitative sensing and imaging. Preprint athttps://doi.org/10.48550/ arXiv.2311.04211(2023)

  45. [53]

    Li, H.et al.Imaging two-dimensional generalized Wigner crystals.Nature597, 650–654 (2021)

  46. [54]

    Wang, T.et al.Probing the edge states of Chern insula- tors using microwave impedance microscopy.Phys. Rev. B108, 235432 (2023)

  47. [55]

    & Kivelson, S

    Spivak, B. & Kivelson, S. A. Phases intermediate be- tween a two-dimensional electron liquid and Wigner crys- tal.Phys. Rev. B70, 155114 (2004)

  48. [56]

    & Changlani, H

    Kumar, A., Lewandowski, C. & Changlani, H. J. Ori- gin and stability of generalized Wigner crystallinity in triangular moir´ e systems.npj Quantum Mater.10, 95 (2025). 10 FIG. 1.AIMS turns three generic experimental uncertainties into action through three nested loops.AIMS con...

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.