Pith. sign in

REVIEW 3 major objections 6 minor 34 references

Tree Models Machine Learning to Identify Liquid Metal based Alloy Superconductor

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read The paper claims that an Extra Trees model trained on one-hot chemical compositions predicts the critical temperature of liquid-metal alloy superconductors with $R^2=0.9519$ and identifies In$_{0.5}$Sn$_{0.5}$ as the highest-$T_c$…

desk verdict A competent screening paper with a modest new application, but the headline In0.5Sn0.5 result is probably retrieval from training data, not a validated prediction. read the letter →

arxiv 2501.05164 v1 pith:J5ORCMFY submitted 2025-01-09 cond-mat.supr-con

classification cond-mat.supr-con
keywords liquidmetalmachinelearningcriticaltemperaturesuperconductorExtraTreesalloyscreeningcompositionencoding
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

Superconducting materials are limited by how hard it is to predict their critical temperature, and the paper tries to show that a fast, formula-only machine-learning model can close that gap for liquid-metal alloys. The authors clean a large literature-derived dataset of known superconductors, train eight tree-based regressors, and report that an Extra Trees model reproduces transition temperatures with $R^2=0.9519$ and an RMSE of 6.2624 K. They then use the model to screen printable liquid-metal alloys, identifying In$_{0.5}$Sn$_{0.5}$ as the best candidate with a predicted $T_c$ of 7.01 K, and extend the screen to thousands of binary and ternary alloys across 66 metals. If the claim holds, alloy discovery for printed superconducting wires and flexible electronics becomes a quick computational ranking instead of slow trial and error.

What carries the argument

The object that carries the argument is the one-hot composition vector: a formula is turned into a long fixed-length vector whose entries are the atomic fractions of each permitted element, so the stoichiometry is the only descriptor. The regression mechanism is Extra Trees, an ensemble of randomized decision trees that averages many trees built with random thresholds, chosen after a grid search among eight tree-based models. The paper argues that tree ensembles are well matched to this tabular regression task because they preserve feature orientation, learn irregular target functions, need little regularization, and train quickly on the roughly 21,000-sample dataset. A five-fold cross-validated hyperparameter search selects the Extra Trees configuration used for all alloy screening.

What would settle it

Synthesize In$_{0.5}$Sn$_{0.5}$ at normal pressure and measure resistivity versus temperature: if the superconducting onset is not near 7.01 K, or if the alloy does not superconduct at all, the composition-only prediction is falsified for the paper's headline candidate. A broader version would measure a sample of the 2,145 binary and 45,760 ternary alloys the model ranks highly and compare predicted and observed $T_c$ values.

Watch

Extended reading notes

Core claim

The paper's central claim is that a material's superconducting critical temperature can be read off its chemical formula: encode the atomic fractions of the elements as a one-hot vector, feed it to an Extra Trees regressor, and the model reproduces known transition temperatures and ranks untested liquid-metal alloys. The authors build this model on a cleaned dataset of 20,995 records, after removing missing, high-pressure, mislabeled, and illegal-element entries, converting mass-percentage formulas to atomic ratios, and adding known non-superconductors with $T_c = 0$. They report the Extra Trees model outperforms seven other tree-based models, reaching $R^2 = 0.9519$ and an RMSE of 6.2624 K on a held-out split. Screening printable liquid-metal alloys from eight elements, the model assigns the highest $T_c$, 7.01 K, to In$_{0.5}$Sn$_{0.5}$, and it is then applied to 2,145 binary and 45,760 ternary alloys across 66 metals. The paper stresses that the model predicts $T_c$ values, not whether a given composition is a superconductor.

Load-bearing premise

The pipeline assumes that the superconducting transition temperature is fully determined by the proportions of the chemical elements, with no influence from crystal structure, pressure, or synthesis conditions; the paper's own discussion notes that one-hot coding records only element proportions and lacks structural information.

Editorial extensions

If this is right

  • Alloy screening becomes a formula-only pipeline: tens of thousands of binary and ternary combinations can be ranked in minutes without density-functional calculations or experiments.
  • Printable liquid-metal electronics gain a specific lead candidate, In$_{0.5}$Sn$_{0.5}$, with a predicted $T_c$ of 7.01 K, worth testing for superconducting wires and flexible circuits.
  • Within a known alloy family, the model can guide composition tuning: the paper reports Ga-In-Sn predictions within about 1 K of experimental values.
  • Tree-based regression is a competitive baseline for $T_c$ prediction from composition, matching or beating neural-network results on this mid-sized dataset.
  • The predicted $T_c$ map over 66 metals gives experimentalists a prioritized list of thousands of plausible alloy superconductors below 40 K.

Reading between the lines

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

  • The paper leaves implicit that the same model could screen quaternary and higher alloys once the combinatorial search becomes efficient; it notes the current search algorithm is the limiting factor.
  • The paper's own limitation statement implies the 7.01 K value is a composition-only estimate; adding crystal-structure or pressure information could change the ranking.
  • A testable follow-up is to check whether prediction residuals cluster in composition neighborhoods, which would reveal families where structure or synthesis conditions override composition.
  • If this recipe transfers beyond superconductors, the one-hot-plus-Extra-Trees pipeline could rank other composition-dependent properties of liquid-metal alloys.
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 / 6 minor

Summary. The paper applies tree-based machine learning regression to the SuperCon database, using a one-hot encoding of chemical composition as the sole descriptor, to predict superconducting critical temperatures (Tc). After preprocessing, the authors compare eight tree-based models and report that an Extra Trees model achieves R2 = 0.9519 and RMSE = 6.2624 K on a random 9:1 train/test split. The model is then used to screen liquid-metal (LM) alloys, identifying In0.5Sn0.5 as the printable LM alloy with the highest predicted Tc at 7.01 K. The authors further screen 2,145 binary and 45,760 ternary alloys formed from 66 metal elements. The central claims are the high predictive accuracy of the Extra Trees model and the identification of In0.5Sn0.5 as a promising high-Tc printable liquid-metal superconductor.

Significance. If the model genuinely generalizes to new compositions, the work would provide a low-cost screening tool for printable liquid-metal superconductors and a large database of predicted Tc values for binary and ternary alloys. The external comparison with Ga-In-Sn experimental data (errors within 1 K) is a useful sanity check, and the systematic comparison of eight tree-based models is a reasonable methodological contribution. However, the headline 'identification' of In0.5Sn0.5 is not established as a predictive discovery: the candidate is a known In-Sn alloy composition that may already exist in the training set, and the one-hot composition descriptor discards structural and pressure information that the authors themselves acknowledge as important. The screening results are thus better described as interpolation/retrieval within the known composition space unless a composition-holdout validation is provided.

major comments (3)
  1. [§3, Fig. 4(c); §2 Data preprocessing] The identification of In0.5Sn0.5 as the highest-Tc printable LM alloy (7.01 K) is not supported as a predictive discovery because the paper never reports whether In_xSn_1-x compositions, including In0.5Sn0.5 itself, appear in the cleaned mdr_clean.csv dataset or in the training partition. Since the model is trained on SuperCon data and In-Sn alloy superconductors are known (ref. 32 reports their synthesis), the 7.01 K value may simply be an interpolation or average of experimental Tc values already in the training set. The authors should perform a composition-based holdout test — for example, training on all data except entries containing In and/or Sn, or at least except In-Sn compositions — and report the predicted Tc for In0.5Sn0.5 under that model. They should also state explicitly whether the exact formula In0.5Sn0.5 is present in the training data. Without this analysis, the central screening claim is unverified.
  2. [§4, Discussion point 1; §2 Model comparison] The one-hot composition descriptor contains only elemental proportions and carries no information about crystal structure, pressure, or synthesis conditions. The authors acknowledge this in the Discussion ('one-hot coding used in this study only records the proportion of elements in the formula. In the absence of structural information'). This is not merely a peripheral limitation: the screening protocol applies the model to arbitrary alloy compositions, including hypothetical ones that may lie far from the training distribution, and the prediction for In0.5Sn0.5 is made under the implicit assumption that composition alone determines Tc. The paper should either restrict the screening claims to compositions within the chemical and structural domain of the training data, or add a similarity/distance check to flag extrapolations. As written, the uncertainty of the 7.01 K prediction is understated.
  3. [§3, Fig. 4(b)] The claim that 1,000 random composition samples per alloy family is sufficient to find the maximum Tc is inferred from the stability of the maximum for one ternary family (Ga-In-Sn). This does not establish that 1k samples is sufficient for all 2,145 binary and 45,760 ternary families, some of which may have sharper Tc optima or more complex composition dependence. A sensitivity analysis on several representative families, or an adaptive sampling criterion with a convergence check, is needed to support the 'highest Tc' claims for the full screening. Without this, the reported maxima for many alloy families may be underestimated.
minor comments (6)
  1. [§3, Eq. (5)] Equation (5) as written has the numerator and denominator of the R2 expression reversed relative to the standard definition R2 = 1 - SS_res/SS_tot. Please correct this typo, since the reported R2 values are otherwise consistent with the standard calculation.
  2. [Abstract vs. §3] The abstract states '45,670 ternary alloys' while Section 3 and Figure 4 report 45,760 ternary alloys; the numbers should be reconciled.
  3. [§5, Conclusion] The conclusion refers to 'mdr_celan.csv' instead of 'mdr_clean.csv'; please correct the typo.
  4. [§3, Eq. (6)-(7)] The quantity in Eq. (6) is the per-sample absolute error, not the mean absolute error (MAE). Rename it (e.g., 'absolute error') or define MAE as the mean over predictions to avoid confusion with the standard metric.
  5. [References] Reference 32 (Mochiku et al., Synthesis of InSn alloy superconductor below room temperature) appears in the reference list but is not cited in the main text. If it is intended to support the existence of In-Sn superconductors in the training set or the relevance of In0.5Sn0.5, it should be cited where appropriate.
  6. [Figure 4] The color maps and element labels in Figures 4(c.ii), 4(c.iii), 4(d.ii), and 4(d.iii) are too small to read in the provided resolution; higher-resolution or separate panels would improve clarity.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the ML pipeline is an independent regression from SuperCon labels; reported limitations concern descriptor information, not derivation self-reference.

full rationale

The paper's derivation chain is a standard supervised machine-learning workflow: SuperCon experimental Tc values serve as labels, one-hot encoded compositions serve as features, an Extra Trees regressor is fit on a 90% training split, and performance is reported on the held-out 10% split. The target Tc is an experimental quantity from the SuperCon database, not a quantity defined in terms of the model's output; hence there is no self-definitional reduction. The reported R2=0.9519 and RMSE=6.2624 are fit-quality metrics on a held-out subset, not fitted parameters renamed as predictions. The headline candidate In0.5Sn0.5 is a model output, and while its reliability would be weakened if that exact composition were present in the training set, the paper does not state this, so a circularity claim would be speculative. The Ga-In-Sn comparison in Figure 5(d) uses values from mdr_clean.csv, but the paper describes a random train/validation split, so those points may legitimately be in the validation set; no explicit in-sample validation is demonstrated. Self-citations (refs 1, 2, 3, 30, 33, 34) appear in background and outlook statements and are not load-bearing for the central claim. The Discussion explicitly flags the one-hot descriptor's lack of structural information, which is a limitation on generalization rather than a circular step. No load-bearing step reduces by construction to its own input, and no external mathematical result is imported from the authors' prior work to force the model choice.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The model's predictions rest on the premise that composition alone determines Tc, that the SuperCon data are taken at face value, that the random sampling search is sufficient, and that the non-superconductor additions are correct. No new physical entities are introduced.

free parameters (3)
  • Extra Trees hyperparameters (n_estimators=300, max_depth=None, min_samples_split=5, min_samples_leaf=1… = n_estimators=300, max_depth=None, min_samples_leaf=1, min_samples_split=5, criterion=friedman_mse
    Selected by grid search with five-fold cross-validation on the cleaned SuperCon dataset (Table 5); the model performance and all downstream predictions depend on this choice.
  • Number of random composition samples per alloy family in screening = 1000
    Chosen after observing that the maximum predicted Tc stabilizes at 7.01 K for one ternary family when sample count exceeds 1000 (Figure 4b); assumed to hold for all 45,760 ternary alloy families without verification.
  • Accuracy threshold (MAE < 5 K) = 5 K
    Defines the reported accuracy metric (Eq. 7); a hand-chosen tolerance that makes the 81.9% accuracy figure look better than a stricter threshold would.
assumptions (4)
  • domain assumption SuperCon dataset entries provide reliable ground-truth Tc values
    The model is trained and evaluated on SuperCon after cleaning; no independent verification of the underlying measurements is performed.
  • domain assumption Tc is determined by elemental composition alone (one-hot formula descriptor)
    All models use one-hot encoded element fractions as the only features; the Discussion acknowledges the absence of structural information, making this a load-bearing premise.
  • domain assumption The 386 non-superconductors with Tc=0 added from Konno and Hosono summaries are correctly labeled
    Non-superconductors are inserted with Tc=0 to improve the dataset, but the labels are taken from summary references rather than the original data, and the model still cannot classify non-superconductors.
  • ad hoc to paper One thousand random composition samples is sufficient to find the maximum Tc for every binary and ternary alloy family
    Figure 4b shows stability of the maximum Tc for one alloy family (Ga-In-Sn) when sampling exceeds 1000; this is extrapolated to all 45,760 ternary families without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Tree Models Machine Learning to Identify Liquid Metal based Alloy Superconductor." pith.science (2026). https://pith.science/paper/J5ORCMFY

@misc{pith2026250105164,
  author       = {Pith},
  title        = {Pith review of: Tree Models Machine Learning to Identify Liquid Metal based Alloy Superconductor},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J5ORCMFY}},
  note         = {Machine review of arXiv:2501.05164}
}
read the original abstract

Superconductors, which are crucial for modern advanced technologies due to their zero-resistance properties, are limited by low Tc and the difficulty of accurate prediction. This article made the initial endeavor to apply machine learning to predict the critical temperature (Tc) of liquid metal (LM) alloy superconductors. Leveraging the SuperCon dataset, which includes extensive superconductor property data, we developed a machine learning model to predict Tc. After addressing data issues through preprocessing, we compared multiple models and found that the Extra Trees model outperformed others with an R2 of 0.9519 and an RMSE of 6.2624 K. This model is subsequently used to predict Tc for LM alloys, revealing In0.5Sn0.5 as having the highest Tc at 7.01 K. Furthermore, we extended the prediction to 2,145 alloys binary and 45,670 ternary alloys across 66 metal elements and promising results were achieved. This work demonstrates the advantages of tree-based models in predicting Tc and would help accelerate the discovery of high-performance LM alloy superconductors in the coming time.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

34 extracted references · 33 canonical work pages

  1. [1]

    Liquid Metal Printed Superconducting Circuits

    Bao, W., Zhang, J., Rao, W. & Liu, J. Liquid metal printed superconducting circuits. Preprint at https://doi.org/10.48550/arXiv.2410.17652 (2024)

  2. [2]

    & Liu, J

    Wang, Q., Y u, Y ., Yang, J. & Liu, J. Fast fabrication of flexible functional circuits based on liquid metal dual-trans printing. Adv. Mater. 27, 7109–7116 (2015)

  3. [3]

    & Liu, J

    Hua, C., Gao, J. & Liu, J. Room temperature self -healing liquid metals: capabilities, applications and challenges. IJSNM 15, 469–501 (2024). 17

  4. [4]

    Oliveira, L. N. Density-functional theory for superconductors. Phys. Rev. Lett. 60, 2430–2433 (1988)

  5. [5]

    Ab initio theory of superconductivity

    Lüders, M. Ab initio theory of superconductivity. I. Density functional formalism and approximate functionals. Phys. Rev. B 72, 024545 (2005)

  6. [6]

    Interactions between electrons and lattice vibrations in a superconductor

    Éliashberg, G. Interactions between electrons and lattice vibrations in a superconductor. J. Exp. Theor. Phys. 11, 696–702 (1960)

  7. [7]

    Neuringer, L. J. & Shapira, Y . Effect of spin-orbit scattering on the upper critical field of high- field superconductors. Phys. Rev. Lett. 17, 81–84 (1966)

  8. [8]

    The superconducting properties of Nb 75Ga25-xAlxA15-type alloys annealed at 800°C for one Week

    Sosnowski, J. The superconducting properties of Nb 75Ga25-xAlxA15-type alloys annealed at 800°C for one Week. Phys. Stat. Sol. (a) 64, 771–775 (1981)

Show all 34 references
  1. [9]

    Konno, T. et al. Deep learning model for finding new superconducto rs. Phys. Rev. B 103, 014509 (2021)

  2. [10]

    Hosono, H. et al. Exploration of new superconductors and functional materials, and fabrication of superconducting tapes and wires of iron pnictides. Science and Technology of Advanced Materials 16, 033503 (2015)

  3. [11]

    Stanev, V . et al. Machine learning modeling of superconducting critical temperature. npj Comput. Mater. 4, 1–14 (2018)

  4. [12]

    A data -driven statistical model for predicting the critical temperature of a superconductor

    Hamidieh, K. A data -driven statistical model for predicting the critical temperature of a superconductor. Comput. Mater. Sci. 154, 346–354 (2018)

  5. [13]

    & Horide, T

    Matsumoto, K. & Horide, T. An acceleration search method of higher Tc superconductors by a machine learning algorithm. Appl. Phys. Express 12, 073003 (2019)

  6. [14]

    & Dordevic, S

    Roter, B. & Dordevic, S. V . Predicting new superconductors and their crit ical temperatures using machine learning. Physica C 575, 1353689 (2020)

  7. [15]

    & Fyta, M

    Díaz Carral, Á., Roitegui, M. & Fyta, M. Interpretably learning the critical temperature of superconductors: Electron concentration and feature dimensionality reduction. APL Materials 12, 041111 (2024)

  8. [16]

    & Alaei, M

    Gashmard, H., Shakeripour, H. & Alaei, M. Predicting superconducting transition temperature through advanced machine learning and innovative feature engineering. Sci. Rep. 14, 3965 (2024)

  9. [17]

    Li, S. et al. Critical temperature prediction of superconductors based on atomic vectors and deep learning. Symmetry 12, 262 (2020)

  10. [18]

    Quinn, M. R. & McQueen, T. M. Identifying new classes of high temperature superconductors with convolutional neural networks. Front. Electron. Mater 2, (2022)

  11. [19]

    & Lawler, K

    Novakovic, L., Salamat, A. & Lawler, K. V . Machine learning using structural representations for discovery of high temperature superconductors. Preprint at https://doi.org/10.48550/arXiv.2301.10474 (2023)

  12. [20]

    Wang, Y . et al. Cuprate superconducting materials above liquid nitrogen temperature from machine learning. RSC Adv. 13, 19836–19845 (2023)

  13. [21]

    Li, J. et al. A deep learning approach to search for superconductors from electronic bands. Preprint at https://doi.org/10.48550/arXiv.2409.07721 (2024)

  14. [22]

    Dan, Y . et al. Computational prediction of critical temperatures of superconductors based on convolutional gradient boosting decision trees. IEEE Access 8, 57868–57878 (2020). 18

  15. [23]

    & Zuo, H

    Hu, J., Jiang, Y ., Yan, Y . & Zuo, H. Prediction of superconducting properties of materials based on machine learning models. Preprint at https://doi.org/10.48550/arXiv.2211.03075 (2022)

  16. [24]

    & Sedaaghi, M

    Taheri, A., Ebrahimnezhad, H. & Sedaaghi, M. H. Prediction of the critical temperature of superconducting materials using image reg ression and ensemble deep learning. Mater. Today Commun. 33, 104743 (2022)

  17. [25]

    & Varoquaux, G

    Grinsztajn, L., Oyallon, E. & Varoquaux, G. Why do tree-based models still outperform deep learning on typical tabular data? in Advances in Neural Information Processing Systems (eds. Koyejo, S. et al.) vol. 35 507–520 (Curran Associates, Inc., 2022)

  18. [26]

    Pedregosa, F. et al. Scikit-learn: machine learning in python. J. Mach. Learn. Res. 12, 2825– 2830 (2011)

  19. [27]

    Prokhorenkova, L., Gusev, G., V orobev, A., Dorogush, A. V . & Gulin, A. CatBoost: unbiased boosting with categorical features. in Proceedings of the 32nd International Conference on Neural Information Processing Systems 6639–6649 (Curran Associates Inc., Red Hook, NY , USA, 2018)

  20. [28]

    & Guestrin, C

    Chen, T. & Guestrin, C. XGBoost: a scalable tree boosting system. in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining 785–794 (Association for Computing Machinery, New York, NY , USA, 2016)

  21. [29]

    Ke, G. et al. LightGBM: a highly efficient gradient boosting decision tree. in Proceedings of the 31st International Conference on Neural Information Processing Systems 3149–3157 (Curran Associates Inc., Red Hook, NY , USA, 2017)

  22. [30]

    & Liu, J

    Zhang, Q., Zheng, Y . & Liu, J. Direct writing of electronics based on alloy and metal (DREAM) ink: A newly emerging area and its impact on energy, environment and health sciences. Front. Energy 6, 311–340 (2012)

  23. [31]

    Ren, L. et al. Nanodroplets for stretchable superconducting circuits. Adv. Funct. Mater. 26, 8111–8118 (2016)

  24. [32]

    & Matsushita, Y

    Mochiku, T., Tachiki, M., Ooi, S. & Matsushita, Y . Synthesis of InSn alloy superconductor below room temperature. Physica C 563, 33–35 (2019)

  25. [33]

    Liquid metal material genome: Initiation of a new research track towards discovery of advanced energy materials

    Wang, L., & Liu, J. Liquid metal material genome: Initiation of a new research track towards discovery of advanced energy materials. Frontiers in Energy 7, 317–332 (2013)

  26. [34]

    Liquid metal combinatorics toward materials discovery

    Wang, D.W., et al. Liquid metal combinatorics toward materials discovery. Adv. Mater. 35, 2303533 (2023)

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.