Pith. sign in

REVIEW 4 major objections 5 minor 44 references

Explainable Multimodal Machine Learning for Revealing Structure-Property Relationships in Carbon Nanotube Fibers

T0 review · 4 major / 5 minor · reviewed 2026-08-08 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Using NMF-extracted dispersion features and SHAP, this paper shows that small, uniformly distributed aggregates improve fracture strength while long effective CNT lengths improve electrical conductivity, with density dominating Young's…

desk verdict A clearly written NMF+SHAP case study on CNT fibers whose central causal claims are undercut by selection-on-the-same-40-samples validation, but which is still worth a serious referee. read the letter →

arxiv 2502.07400 v1 pith:4VTQD3DM submitted 2025-02-11 cond-mat.mtrl-sci cond-mat.softcs.AIcs.LGphysics.data-an

Daisuke Kimura , Naoko Tajima , Toshiya Okazaki , Shun Muroga This is my paper · ORCID
classification cond-mat.mtrl-scicond-mat.softcs.AIcs.LGphysics.data-an
keywords carbonnanotubesCNTfibersexplainableAISHAPnonnegativematrixfactorizationdispersionstatestructure-propertyrelationshipsmultimodalmachinelearning
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

The paper proposes an Explainable Multimodal Machine Learning (EMML) pipeline that combines multiscale characterization, nonnegative matrix factorization (NMF) for feature extraction, Random Forest regression, and SHAP analysis to identify which structural features of carbon nanotube (CNT) fibers actually drive macroscopic properties. Applied to 40 aqueous-dispersion CNT fibers, it finds that fracture strength is governed by small, uniformly sized aggregates in the dispersion, electrical conductivity by long effective CNT lengths and a narrow aggregate-size distribution, and Young's modulus by fiber density and morphology. The claim matters because it converts opaque structure-property correlations into concrete processing targets for bottom-up nanomaterial manufacturing.

What carries the argument

NMF (nonnegative matrix factorization) converts each DCS aggregate-size distribution into weights on seven nonnegative basis distributions, each with a clear peak position and width grouping (small, medium, large). These basis weights are the interpretable features that Random Forest regression consumes along with FIR effective length, Raman I_G/I_D, density, linear density, and process variables. SHAP interaction values then decompose each prediction into contributions per feature, yielding both global importance rankings and local threshold/trend relationships. The paper's methodological claim is that for nonnegative, multi-peaked distribution data, NMF yields sparser and more physically interpretable components than PCA, which is why the DCS-derived basis weights can carry mechanistic meaning.

What would settle it

Prepare fibers from dispersions whose DCS aggregate-size distributions are deliberately varied (for example, by controlled sonication or filtration) while holding other variables fixed, and check whether fracture strength and conductivity track the SHAP-identified basis weights; alternatively, bootstrap the 40-sample dataset to see whether the same seven basis components and their SHAP signs reappear.

Watch

Extended reading notes

Core claim

The paper's own claim is that the structure-property relationships in aqueous-dispersion CNT fibers can be read out of a multimodal dataset by decomposing aggregation-size distribution data with NMF into seven interpretable basis distributions, then applying SHAP to a Random Forest trained on those basis weights plus effective CNT length, Raman I_G/I_D, density, linear density, and processing conditions. On the 40 measured fibers, fracture strength is controlled by small, uniformly distributed aggregates (DCS Basis000 and Basis001) and a small amount of medium-sized aggregates (Basis004), while even a slight presence of large aggregates (Basis006) sharply lowers strength; electrical conductivity requires long effective CNT length (with a threshold near 2000 nm) and a narrow distribution of small aggregates; and Young's modulus is governed by macroscale density and fiber morphology rather than dispersion features. The authors interpret this as evidence that aggregate assemblies, not individual CNTs, are the fundamental structural unit determining CNT fiber properties.

Load-bearing premise

The central attribution depends on the seven-component NMF decomposition, chosen by maximizing leave-one-out prediction accuracy on the same 40 samples, being a stable and physically meaningful representation of the dispersion state rather than an overfit artifact.

Editorial extensions

If this is right

  • Fracture strength can be pursued by engineering dispersions with a narrow size distribution centered on small aggregates and only a minor medium-sized component, because these features maximize aggregate-aggregate junctions for load transfer.
  • Electrical conductivity can be pursued by lengthening CNTs and keeping their effective length above roughly 2000 nm while minimizing the broadening component of the small-aggregate distribution.
  • Young's modulus optimization should concentrate on spinning conditions that raise fiber density and reduce void content, since dispersion-state features have minor influence on stiffness.
  • The same EMML workflow is transferable to other nanomaterial-derived materials where multiscale structures and distribution-type measurements obscure which processing step matters.

Reading between the lines

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

  • A direct testable extension would be to check the stability of the seven NMF bases under bootstrap resampling of the 40 samples; if the bases change, the physical interpretation of Basis000/001/004/006 would need re-basing on independent data.
  • The paper leaves implicit that the identified thresholds (effective length around 2000 nm, I_G/I_D around 30) are dataset-specific; testing whether they survive across different CNT synthesis routes or dispersants would establish their generality.
  • Because density is the dominant factor for Young's modulus while dispersion dominates strength, the results imply competing processing priorities: densification may tighten packing but could disturb the narrow aggregate distribution, so multi-objective optimization is needed.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes an Explainable Multimodal Machine Learning (EMML) pipeline for CNT fibers, combining multimodal characterization data (DCS aggregate size distributions, FIR effective length, Raman IG/ID, morphological features, and processing conditions) with NMF-based feature extraction, random forest regression, and SHAP-based interpretation. Using 40 CNT fiber samples from a prior experimental study, the authors select a 7-component NMF decomposition of DCS distributions based on LOOCV R2, then report SHAP global importance and local interaction analyses. They conclude that fracture strength is improved by small, uniformly distributed aggregates and a small amount of medium-sized aggregates with wide distribution widths; electrical conductivity is enhanced by long effective CNT lengths and a narrow distribution of small aggregates; and Young's modulus is dominated by density and fiber morphology. Thresholds for IG/ID and effective length are also proposed as processing targets.

Significance. If the results are robust, the paper offers a useful template for combining factor analysis with explainable AI in multiscale materials informatics, and it connects specific, measurable dispersion-state features (NMF basis weights) to macroscopic fiber properties. The comparison of PCA and NMF on the DCS distributions is a clear strength, and the local SHAP interaction plots provide more actionable information than standard feature-importance rankings. The claimed structure-property relationships are plausible and consistent with prior CNT fiber literature, and the proposed thresholds could guide dispersion-process optimization. However, the significance is conditional on the stability and independent validation of the 7-component NMF model and the random forest regressions; the current evidence is not yet sufficient to support the strong mechanistic language used in the abstract and Section 3.4.

major comments (4)
  1. [Section 2; Section 3.2; Figure 3(a); Figure S2(c)] The number of NMF components and the random forest hyperparameters are both selected by maximizing LOOCV R2 on the same 40 samples, and the reported R2 values and SHAP attributions are then derived from that same selected model. This procedure leaks information from the left-out folds into the model-selection step, so the LOOCV R2 in Figure 3(a) and the corresponding actual-versus-predicted plots likely overstate predictive performance. Because the central claim that the 7-component decomposition yields meaningful aggregate populations depends on this selection, the authors should either perform a nested cross-validation (selecting NMF rank and hyperparameters within each training fold) or provide an external test set, and they should report confidence intervals for the R2 values and for the NMF component weights across resamples.
  2. [Section 3.2; Section 3.3; Figures 4 and 8] NMF solutions are not unique: they are subject to scaling and component-ordering ambiguities and can depend on the initialization and on small perturbations of the input data. The paper does not report any stability analysis, such as running NMF with multiple random seeds, bootstrapping the 40 DCS distributions, or clustering the resulting basis vectors across runs. Since the interpretation in Figure 8 assigns distinct physical meanings to Basis000, Basis001, Basis004, and Basis006, the authors need to demonstrate that these components are stable under resampling and re-initialization; otherwise the SHAP-based attribution of strength and conductivity changes to specific basis components could be an artifact of a particular factorization.
  3. [Section 3.3; Figures 7 and 8] The claim that SHAP interaction values should be interpreted only above a "threshold of 10%" is introduced without a formal justification, and the threshold is applied after scaling the left axis to approximately 5% of the maximum target value. This is load-bearing for the threshold statements about IG/ID ~30 and effective length ~2000 nm, because different scaling choices could change which points fall above the threshold. The authors should either derive the threshold from a principled criterion (for example, a noise floor estimated from repeated model fits) or present the sensitivity of the qualitative conclusions to the chosen threshold.
  4. [Section 3.4; Figure 9] The language in the abstract and conclusion states that small, uniformly distributed aggregates are 'crucial' for improving fracture strength and that long effective lengths are 'significant factors' for conductivity, implying causal or mechanistic status. What the analysis actually establishes is predictive association within one set of 40 samples. The authors themselves note that XAI 'does not directly clarify mechanisms' in Section 2, but the later discussion nonetheless draws causal conclusions such as 'the presence of these aggregates introduces subtle structural heterogeneity, which helps to reduce crack and promote the redistribution of stress.' These mechanistic interpretations should be clearly flagged as hypotheses consistent with the SHAP analysis, not as demonstrated mechanisms, unless supported by targeted experiments or physical simulations.
minor comments (5)
  1. [Abstract; Introduction] The term 'Negative Matrix Factorization' is used in the Abstract; the correct standard term is 'Non-negative Matrix Factorization' (also used elsewhere in the paper).
  2. [Figure 7 caption and Section 3.3] The text discusses the IG/ID ratio, but the right-axis labels in Figure 7 (and the corresponding supplementary figures S10-S12) show 'ID / IG'. Please make the notation consistent.
  3. [Section 3.2, Figure 3(a)] The text states that the coefficient of determination reached its maximum at 7 components for all three target variables, but the figure shows only the general trend; please indicate explicitly whether the maximum is attained at exactly 7 for each property or whether 7 is a compromise across the three targets.
  4. [Section 3.3, Figure 5] The sentence 'Figure 5 provides a global evaluation of how explanatory variables to regression predictions' appears to be missing a verb (e.g., 'contribute to'); please rephrase.
  5. [Supplementary Figures S10-S12] The unit 'mg/cm' is used for linear density, which is acceptable if intended as tex-like linear density, but it should be defined clearly in the main text or captions to avoid confusion with volumetric density.

Circularity Check

2 steps flagged · score 5.0 of 10

The 'discovered' structure-property relationships are read from a model whose NMF rank and hyperparameters were selected to maximize the same LOOCV R2 on the same 40 samples, making the SHAP attributions partly selection artifacts rather than independent derivations.

  1. fitted input called prediction [Section 2 (Material and methods), Section 3.2, Figure 3(a)]
    "Furthermore, the number of factors in NMF was optimized based on predictive performance (Figure S2). ... For fracture strength, electrical conductivity, and Young's modulus, the coefficient of determination reached its maximum when the number of components was set to 7."

    The NMF component count is selected by maximizing the LOOCV coefficient of determination for the same target variables (fracture strength, electrical conductivity, Young's modulus) that are later 'predicted' and interpreted via SHAP. Since the same 40 samples are used both to choose the number of factors and to compute the reported LOOCV R2 and feature importances, the selected Basis000/Basis001/Basis004/Basis006 features are not independent of the targets: they are the components that best fit those targets. The subsequent attribution of mechanical and electrical roles to these components is therefore partly a restatement of the selection criterion, not a standalone structural derivation.

  2. fitted input called prediction [Supplementary Figure S2(c); Section 2]
    "The optimal hyperparameters were determined based on the highest coefficient of determination (R2). ... Finally, SHAP values were calculated for the model trained under optimized conditions."

    Random forest hyperparameters are tuned on the same Leave-One-Out R2 used to report the model's predictive reliability, and the SHAP values are computed from that same optimized model. Any LOOCV R2 and SHAP pattern reported after this tuning is optimistically biased, and the 'demonstrated' contributions of effective length and DCS basis features are post hoc descriptions of a model fitted to the same 40 samples, with no external or hold-out validation to show they are stable rather than selected.

full rationale

The derivation chain is not definitionally circular: NMF decomposition is applied to measured DCS distributions before the target variables are involved, and SHAP does quantify the actual fitted model's behavior. A LOOCV R2 also gives a genuine, if noisy, estimate of internal predictive accuracy. However, the central structure-property 'discoveries' are not independent of their inputs. The number of NMF components is chosen by maximizing the same LOOCV R2 for the same three target properties, and the random forest hyperparameters are tuned on the same R2, all on the same 40 samples; the SHAP values then describe that target-selected model. Thus the reported importance of small/uniform aggregates and long effective lengths is partly a selection artifact: those features were preferentially chosen because they correlated with the targets. The paper contains no external test set or stability analysis to separate genuine relationships from overfitting to these 40 samples. This is a partial circularity and selection-bias problem, not an equivalence-by-definition, and self-citations to prior group work are not load-bearing.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The central claims rest primarily on the fidelity of DCS-derived aggregate distributions, the target-informed choice of 7 NMF components, and the assumption that SHAP on a 40-sample Random Forest reflects real structure-property relationships. No new physical entities are introduced; the NMF basis components are latent data features, not postulated physical objects. The free parameters are the NMF rank, the random forest hyperparameters, and the SHAP meaningful-contribution threshold.

free parameters (3)
  • Number of NMF components = 7
    Selected by maximizing LOOCV R2 for all three target properties in Section 3.2, Figure 3(a). This target-informed selection makes the seven basis features partly fitted to the outcome variables.
  • Random Forest hyperparameters = not reported
    Hyperparameter tuning is mentioned in Figure S2(c), but the actual values and search ranges are not given. The final model and SHAP values depend on these choices.
  • SHAP meaningful-contribution threshold = 10%
    Adopted in Section 3.3 to declare a SHAP interaction contribution meaningful. The threshold is arbitrary and not justified by any statistical test.
assumptions (5)
  • standard math Non-negative matrix factorization and SHAP are mathematically well-defined for the input data.
    The analysis relies on standard NMF decompositions and SHAP value computations; these are accepted mathematical tools used without formal proof in the paper.
  • domain assumption DCS aggregation size distributions faithfully represent the dispersion state relevant to final fiber properties.
    All DCS-based conclusions in Sections 3.2 and 3.3 assume this measurement captures the controlling structural feature of the dispersion.
  • domain assumption A 7-component NMF decomposition corresponds to physically meaningful aggregate classes.
    The interpretability of Basis000-006 as small, medium, and large aggregate populations is asserted in Section 3.2 but not independently validated against microscopy or another structural probe.
  • domain assumption SHAP values on the trained Random Forest approximate causal contributions of structural features.
    The paper itself states that XAI does not directly clarify mechanisms in Section 2, yet Section 3.4 presents the SHAP-based findings as mechanism elucidation.
  • domain assumption Leave-one-out cross-validation with target-informed hyperparameter and factor selection yields unbiased predictive performance.
    LOOCV after selecting NMF rank and hyperparameters on the same data can be optimistic; no independent held-out test set is used.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Explainable Multimodal Machine Learning for Revealing Structure-Property Relationships in Carbon Nanotube Fibers." pith.science (2026). https://pith.science/paper/4VTQD3DM

@misc{pith2026250207400,
  author       = {Pith},
  title        = {Pith review of: Explainable Multimodal Machine Learning for Revealing Structure-Property Relationships in Carbon Nanotube Fibers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4VTQD3DM}},
  note         = {Machine review of arXiv:2502.07400}
}
read the original abstract

In this study, we propose Explainable Multimodal Machine Learning (EMML), which integrates the analysis of diverse data types (multimodal data) using factor analysis for feature extraction with Explainable AI (XAI), for carbon nanotube (CNT) fibers prepared from aqueous dispersions. This method is a powerful approach to elucidate the mechanisms governing material properties, where multi-stage fabrication conditions and multiscale structures have complex influences. Thus, in our case, this approach helps us understand how different processing steps and structures at various scales impact the final properties of CNT fibers. The analysis targeted structures ranging from the nanoscale to the macroscale, including aggregation size distributions of CNT dispersions and the effective length of CNTs. Furthermore, because some types of data were difficult to interpret using standard methods, challenging-to-interpret distribution data were analyzed using Negative Matrix Factorization (NMF) for extracting key features that determine the outcome. Contribution analysis with SHapley Additive exPlanations (SHAP) demonstrated that small, uniformly distributed aggregates are crucial for improving fracture strength, while CNTs with long effective lengths are significant factors for enhancing electrical conductivity. The analysis also identified thresholds and trends for these key factors to assist in defining the conditions needed to optimize CNT fiber properties. EMML is not limited to CNT fibers but can be applied to the design of other materials derived from nanomaterials, making it a useful tool for developing a wide range of advanced materials. This approach provides a foundation for advancing data-driven materials research.

Figures

Figures reproduced from arXiv: 2502.07400 by the authors.

Figure 1
Figure 1. Overview of property predictions utilizing machine learning and explainable AI to support mechanism understanding. Fibers Numerical Vector Textual … … Fracture strength Electrical conductivity Young’s modulus Process and Characteristic Property AI model Actual Predicted Explainable AI Importance Relationship Slurry CNTs Mechanism Understanding Materials design [PITH_FULL_IMAGE:figures/full_fig_p012_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 39 canonical work pages

  1. [1]

    These fibers feature a structure in which CNTs, as nanoscale materials, assemble into bundles

    Introduction Carbon nanotube (CNT) fibers [1 –12] have garnered attention as next -generation materials due to their exceptional mechanical, electrical, and thermal properties. These fibers feature a structure in which CNTs, as nanoscale materials, assemble into bundles. There is great interest in transferring the remarkable properties of individual CNTs ...

    work page

  2. [2]

    Material and methods The dataset in this study is derived from CNT fibers prepared from aqueous dispersions reported by Tajima et al. [10]. The preparation of CNT fibers using aqueous dispersions has garnered attention as a highly safe and sustainable method, with several studies having been reported. The dispersions used in this study were prepared using...

    work page

  3. [3]

    To accurately investigate these hierarchical structures, advanced analytical techniques specialized for CNTs have been developed

    Results and discussion 3.1 Multiscale Structural Analysis Techniques and Factor Analysis of Distribution Data CNT fibers exhibit multiscale structures, with features ranging from nanoscale defects and aggregation structures to macroscale formations. To accurately investigate these hierarchical structures, advanced analytical techniques specialized for CNT...

    work page 2000

  4. [4]

    EMML integrate d multimodal data analysis through factor analysis and enhances interpretability using Explainable AI (XAI)

    Conclusion In conclusion, we applied EMML (Explainable Machine Learning) to analyze CNT fibers fabricated from aqueous dispersions, aiming to elucidate the complex mechanisms governing their material properties under multi-stage fabrication conditions and multiscale structural influences. EMML integrate d multimodal data analysis through factor analysis a...

    work page

  5. [5]

    S. G. Kim et al., Advanced Science, 9 (2022) 2203008. (https://doi.org/10.1002/advs.202203008)

    work page doi:10.1002/advs.202203008 2022

  6. [6]

    L. W. Taylor et al., Carbon, 171 (2021) 689. (https://doi.org/10.1016/j.carbon.2020.07.058)

    work page doi:10.1016/j.carbon.2020.07.058 2021

  7. [7]

    Poulin et al., Carbon, 40 (2002) 1741

    P. Poulin et al., Carbon, 40 (2002) 1741. (https://doi.org/10.1016/S0008-6223(02)00042-8)

    work page doi:10.1016/s0008-6223(02)00042-8 2002

  8. [8]

    Koiol et al., Science, 318 (2007) 1892

    K. Koiol et al., Science, 318 (2007) 1892. (https://doi.org/10.1126/science.1147635)

    work page doi:10.1126/science.1147635 2007

Show all 44 references
  1. [9]

    Vigolo et al., Science, 290 (2000) 1331

    B. Vigolo et al., Science, 290 (2000) 1331. (https://doi.org/10.1126/science.290.5495.1331)

    2000

  2. [10]

    Tajima et al., Carbon, 152 (2019) 1

    N. Tajima et al., Carbon, 152 (2019) 1. (10.1016/j.carbon.2019.05.062)

    2019 doi

  3. [11]

    Behabtu et al., Science, 339 (2013) 182

    N. Behabtu et al., Science, 339 (2013) 182. (https://doi.org/10.1126/science.1228061)

    2013 doi

  4. [12]

    Lekawa-Raus et al., Advanced Functional Materials, 24 (2014) 3661

    A. Lekawa-Raus et al., Advanced Functional Materials, 24 (2014) 3661. (https://doi.org/10.1002/adfm.201303716)

    2014 doi

  5. [13]

    Wu et al., Applied Physics Express, 10 (2017) 055101

    X. Wu et al., Applied Physics Express, 10 (2017) 055101. (https://doi.org/10.7567/APEX.10.055101)

    2017 doi

  6. [14]

    Wu et al., The Journal of Physical Chemistry C, 120 (2016) 20419

    X. Wu et al., The Journal of Physical Chemistry C, 120 (2016) 20419. (https://doi.org/10.1021/acs.jpcc.6b06746)

    2016 doi

  7. [15]

    T. W. Tombler et al., Nature, 405 (2000) 769. (https://doi.org/10.1038/35015519)

    2000 doi

  8. [16]

    B. Q. Wei et al., Applied Physics Letters, 79 (2001) 1172. (https://doi.org/10.1063/1.1396632)

    2001 doi

  9. [17]

    P. L. McEuen et al., IEEE Transactions on Nanotechnology, 1 (2002) 78. (https://doi.org/10.1109/TNANO.2002.1005429)

    2002 arXiv

  10. [18]

    Avouris et al., Proceedings of the IEEE, 91 (2003) 1772

    P. Avouris et al., Proceedings of the IEEE, 91 (2003) 1772. (https://doi.org/10.1109/JPROC.2003.818338)

    2003

  11. [19]

    Mir et al., Applied Mathematical Modeling, 45 (2017) 1031

    M. Mir et al., Applied Mathematical Modeling, 45 (2017) 1031. (https://doi.org/10.1016/j.apm.2017.01.038)

    2017 doi

  12. [20]

    Akcora et al., Nature Materials, 8 (2009) 354

    P. Akcora et al., Nature Materials, 8 (2009) 354. (https://doi.org/10.1038/nmat2404)

    2009 doi

  13. [21]

    B. G. Demczyk et al., Materials Science and Engineering: A, 334 (2002) 173. (https://doi.org/10.1016/S0921- 5093(01)01807-X)

    2002 doi

  14. [22]

    M. F. Yu et al., Physical Review Letters, 84 (2000) 5552. (https://doi.org/10.1103/PhysRevLett.84.5552)

    2000 doi

  15. [23]

    M. F. Yu et al., Science, 287 (2000) 637. (https://doi.org/10.1126/science.287.5453.637)

    2000 doi

  16. [24]

    Bai et al., Nature Nanotechnology, 13 (2018) 589

    Y. Bai et al., Nature Nanotechnology, 13 (2018) 589. (https://doi.org/10.1038/s41565-018-0141-z)

    2018 doi

  17. [25]

    Nonoguchi et al., Advanced Materials Interfaces, 9 (2022) 2101723

    Y . Nonoguchi et al., Advanced Materials Interfaces, 9 (2022) 2101723. (https://doi.org/10.1002/admi.202101723)

    2022 doi

  18. [26]

    Bouville and A

    F. Bouville and A. R. Studart, Nature Communications, 8 (2017) 14655. (https://doi.org/10.1038/ncomms14655)

    2017 doi

  19. [27]

    Dong et al., Advanced Functional Materials, 31 (2021) 2007750

    Y . Dong et al., Advanced Functional Materials, 31 (2021) 2007750. (https://doi.org/10.1002/adfm.202007750)

    2021 doi

  20. [28]

    Qin et al., Advanced Functional Materials, 20 (2020) 2003429

    X. Qin et al., Advanced Functional Materials, 20 (2020) 2003429. (https://doi.org/10.1002/adfm.202003429)

    2020 doi

  21. [29]

    M. H. Bistervels et al., Advanced Materials, 34 (2022) 2107843. (https://doi.org/10.1002/adma.202107843)

    2022 doi

  22. [30]

    Fu et al., Carbon, 223 (2024) 119045

    Q. Fu et al., Carbon, 223 (2024) 119045. (https://doi.org/10.1016/j.carbon.2024.119045)

    2024

  23. [31]

    Jintoku and D

    H. Jintoku and D. N. Futaba, ACS Applied Materials & Interfaces, 16 (2024) 111800. (https://doi.org/10.1021/acsami.3c18612)

    2024 doi

  24. [32]

    Lin et al., ACS Nano, 17 (2023) 22821

    D. Lin et al., ACS Nano, 17 (2023) 22821. (https://doi.org/10.1021/acsnano.3c07587)

    2023 doi

  25. [33]

    Muroga et al., Advanced Science, 10 (2023) 2302508

    S. Muroga et al., Advanced Science, 10 (2023) 2302508. (https://doi.org/10.1002/advs.202302508)

    2023 doi

  26. [34]

    M. I. M. Kusdhany and S. M. Lyth, Carbon, 179 (2021) 190. (https://doi.org/10.1016/j.carbon.2021.04.036)

    2021 doi

  27. [35]

    M. S. Dresselhaus et al., Carbon, 40 (2002) 2043 (https://doi.org/10.1016/S0008-6223(02)00066-0)

    2002 doi

  28. [36]

    M. A. Haque et al., Materials Today Communications, 35 (2023) 105547. (https://doi.org/10.1016/j.mtcomm.2023.105547)

    2023

  29. [37]

    for feature extraction from distribution data. PCA is one of the most fundamental and widely used dimensionality reduction techniques, which efficiently projects high-dimensional data onto a lower-dimensional space based on orthogonal bases that maximizes data variance. In con...

  30. [38]

    S. M. Lundberg et al., Nature Machine Intelligence, 2 (2020) 56. (https://doi.org/10.1038/s42256-019-0138-9)

    2020 doi

  31. [39]

    Nadler et al., Carbon, 46 (2008) 1384

    M. Nadler et al., Carbon, 46 (2008) 1384. (https://doi.org/10.1016/j.carbon.2008.05.024)

    2008 doi

  32. [40]

    Morimoto et al., ACS Nano, 8 (2014) 9897

    T. Morimoto et al., ACS Nano, 8 (2014) 9897. (https://doi.org/10.1021/nn505430s)

    2014 doi

  33. [41]

    E. F. Antunes et al., Carbon, 45 (2007) 913. (https://doi.org/10.1016/j.carbon.2007.01.003)

    2007 doi

  34. [42]

    D. D. Lee and H. S. Seung, Nature, 401 (1999) 788. (https://doi.org/10.1038/44565)

    1999 doi

  35. [43]

    Im et al., Carbon, 232 (2025) 119766

    J. Im et al., Carbon, 232 (2025) 119766. (https://doi.org/10.1016/j.carbon.2024.119766)

    2025

  36. [44]

    Weibull, Journal of Applied Mechanics, 18 (1951) 293

    W. Weibull, Journal of Applied Mechanics, 18 (1951) 293. (https://doi.org/10.1115/1.4010337) Acknowledgments We appreciate Dr. Don N. Futaba for his advice about this paper. CRediT authorship contribution statement Conceptualization: S.M., T.O., Methodology: D. K. S.M., Softwa...

    1951 doi

Pith tools

Reviewed August 8, 2026 · model on record in the stance chip above.