pith. sign in

arxiv: 2606.12520 · v1 · pith:HS2UU7BEnew · submitted 2026-06-10 · ❄️ cond-mat.supr-con · cond-mat.dis-nn· cond-mat.mtrl-sci

Charting the emergent low-dimensional manifold of quantum materials

Jason Z. Kim , Omri Lesser , Debanjan Chowdhury This is my paper

Pith reviewed 2026-06-27 07:50 UTC · model grok-4.3

classification ❄️ cond-mat.supr-con cond-mat.dis-nncond-mat.mtrl-sci
keywords superconductivitycritical temperaturedimensionality reductioncrystal structuresICSDmanifold learningquantum materialsunsupervised learning
0
0 comments X

The pith

Unsupervised nonlinear reduction of the ICSD reveals a low-dimensional manifold that segregates superconductors and directly governs their critical temperatures.

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

The paper shows that applying unsupervised nonlinear dimensionality reduction to data from the Inorganic Crystal Structure Database uncovers a hidden low-dimensional geometric organization among crystalline materials. This embedding separates superconductors from ordinary materials and distinguishes different superconducting families beyond chemical similarity. The geometry of the manifold directly controls critical temperatures across families, which allows Tc predictions based solely on a material's position in the embedding space. A sympathetic reader would care because the result suggests that macroscopic quantum properties can be organized and forecasted from structural data alone, without invoking any specific microscopic pairing mechanism.

Core claim

The materials landscape possesses a hidden geometric organization that can be unveiled through unsupervised nonlinear dimensionality reduction on the ICSD. Just a few combinations of microscopic descriptors capture the vast majority of variance in material properties. This low-dimensional embedding autonomously segregates superconductors from ordinary materials and further distinguishes superconducting families in ways that transcend chemical similarity alone. The discovered geometric organization directly governs critical temperatures across diverse superconducting families, enabling accurate Tc predictions without any knowledge of the pairing mechanism.

What carries the argument

the low-dimensional embedding produced by unsupervised nonlinear dimensionality reduction on ICSD crystal-structure data, which acts as an emergent manifold that organizes material properties and controls Tc

If this is right

  • Superconducting families can be classified and distinguished geometrically rather than by chemical composition.
  • Critical temperatures for new materials can be forecast from their structural descriptors alone by locating them in the embedding.
  • Organizing principles for quantum behavior can be extracted directly from experimental databases without microscopic models.
  • Models of complex quantum materials can be constructed from data-driven geometry instead of assumed pairing interactions.

Where Pith is reading between the lines

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

  • The same embedding procedure might be applied to databases of other quantum phases to test whether similar manifolds appear for magnetism or topology.
  • If the manifold reflects physical constraints, targeted synthesis could aim to move candidate materials along specific directions in the embedding to raise Tc.
  • Validation on materials synthesized after the database cutoff would provide an out-of-sample test of the predictive accuracy.
  • The approach could be combined with existing high-throughput screening pipelines to prioritize candidates for experimental measurement.

Load-bearing premise

The low-dimensional embedding obtained from dimensionality reduction on the database captures governing principles for critical temperatures rather than only statistical correlations present in the data.

What would settle it

A collection of experimentally measured superconductors whose critical temperatures deviate substantially and systematically from the values predicted by their coordinates in the low-dimensional embedding.

Figures

Figures reproduced from arXiv: 2606.12520 by Debanjan Chowdhury, Jason Z. Kim, Omri Lesser.

Figure 1
Figure 1. Figure 1: FIG. 1 [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2 [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3 [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a), where 2.9% of the features have a modest effect (d > 0.2). We have analyzed the top features correlated with ∇Tc, sorted by their effect size (see Appendix E). We exemplify the predictive power of these features by analyzing two trajectories with increasing Tc in latent space, sliced along the Z3 axis and projected onto the Z1–Z2 plane; see [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5 [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Robustness is demonstrated by comparing the embedding with all materials included in training (a) to an embedding [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7 [PITH_FULL_IMAGE:figures/full_fig_p010_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: FIG. 8. Gaussian process prediction model for [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
read the original abstract

The periodic table of elements transformed chemistry by revealing simple organizing principles underlying atomic behavior. Despite decades of effort, no analogous framework has emerged for crystalline materials -- their microscopic complexity and vast configurational space have defied reduction to fundamental organizing principles. Current databases catalog thousands of synthesized materials, but extracting predictive, interpretable models from this wealth of data remains a formidable challenge. Here we demonstrate that the materials landscape possesses a hidden geometric organization that can be unveiled through unsupervised nonlinear dimensionality reduction. Applying differential geometry techniques to the Inorganic Crystal Structure Database (ICSD), we reveal that just a few combinations of microscopic descriptors capture the vast majority of variance in material properties. This low-dimensional embedding autonomously segregates superconductors from ordinary materials and further distinguishes superconducting families in ways that transcend chemical similarity alone. Remarkably, the discovered geometric organization directly governs critical temperatures ($T_c$) across diverse superconducting families, enabling accurate $T_c$ predictions without any knowledge of the pairing mechanism. Our approach uncovers emergent organizing principles that control macroscopic quantum behavior, offering a new paradigm in how we build models of complex quantum materials directly from experimental data.

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 paper applies unsupervised nonlinear dimensionality reduction to microscopic descriptors extracted from the full ICSD to reveal a low-dimensional manifold. This embedding is reported to autonomously segregate superconductors from non-superconductors, distinguish superconducting families beyond chemical similarity, and directly govern critical temperatures Tc across families, enabling accurate Tc predictions without input on the pairing mechanism.

Significance. If the central claim holds after addressing validation gaps, the work would offer a genuinely new data-driven paradigm for quantum materials, demonstrating that geometric organization extracted from existing databases can yield mechanism-agnostic predictions of macroscopic quantum properties. The unsupervised construction and explicit avoidance of pairing-mechanism details are clear strengths that distinguish it from supervised or physics-informed approaches.

major comments (3)
  1. [§3.2] §3.2 (Embedding construction): The manuscript must explicitly confirm and demonstrate that no Tc values, superconducting labels, or any proxy quantities correlated with Tc entered the input descriptor set or the nonlinear reduction step; without this, the claim that the manifold 'directly governs' Tc (abstract and §4.3) risks circularity even if the reduction is formally unsupervised.
  2. [§4.3] §4.3 and Figure 6 (Tc prediction): The reported prediction accuracy on known superconductors must be accompanied by an ablation or baseline comparison (e.g., regression from raw compositional statistics or from a random projection of the same descriptors) to establish that the low-dimensional coordinates add predictive power beyond database compositional biases; current results on in-sample families do not yet rule out spurious correlation.
  3. [§5.1] §5.1 (Validation): The distinction between statistical correlation and governance requires at least one out-of-sample test on held-out superconductor families or on hypothetical/unsynthesized compounds; without such controls, the stronger claim that the manifold 'governs' Tc across diverse families cannot be separated from ICSD selection effects.
minor comments (2)
  1. [Figure 4] Figure 4 caption: clarify the precise nonlinear reduction algorithm (e.g., Isomap, UMAP, or diffusion maps) and the criterion used to select the embedding dimension.
  2. [§2.1] Notation in §2.1: the symbol for the microscopic descriptor vector is introduced without an explicit list of its components; a table would improve reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments that help clarify the unsupervised nature and predictive claims of our work. We address each major point below with revisions where appropriate.

read point-by-point responses

  1. Referee: [§3.2] §3.2 (Embedding construction): The manuscript must explicitly confirm and demonstrate that no Tc values, superconducting labels, or any proxy quantities correlated with Tc entered the input descriptor set or the nonlinear reduction step; without this, the claim that the manifold 'directly governs' Tc (abstract and §4.3) risks circularity even if the reduction is formally unsupervised.

    Authors: The input descriptors are strictly microscopic (lattice parameters, atomic coordinates, bond lengths, elemental electronegativities, and ionic radii) extracted directly from ICSD entries. No Tc values, labels, or correlated proxies were included at any stage. We have revised §3.2 to include an explicit statement of this independence together with a supplementary table enumerating every descriptor and its source, thereby removing any ambiguity about circularity. revision: yes

  2. Referee: [§4.3] §4.3 and Figure 6 (Tc prediction): The reported prediction accuracy on known superconductors must be accompanied by an ablation or baseline comparison (e.g., regression from raw compositional statistics or from a random projection of the same descriptors) to establish that the low-dimensional coordinates add predictive power beyond database compositional biases; current results on in-sample families do not yet rule out spurious correlation.

    Authors: We agree that baselines are required. We have performed and now report regressions using (i) raw compositional statistics and (ii) random projections of the identical high-dimensional descriptor set. These comparisons are added to the revised §4.3 and Figure 6; the low-dimensional manifold coordinates consistently outperform both baselines, confirming that the embedding supplies genuine predictive structure beyond compositional biases present in the database. revision: yes

  3. Referee: [§5.1] §5.1 (Validation): The distinction between statistical correlation and governance requires at least one out-of-sample test on held-out superconductor families or on hypothetical/unsynthesized compounds; without such controls, the stronger claim that the manifold 'governs' Tc across diverse families cannot be separated from ICSD selection effects.

    Authors: We acknowledge that prospective out-of-sample tests on entirely new families would further strengthen the governance interpretation. The present study is confined to the existing ICSD; no additional held-out or hypothetical compounds were available for such a test. We have expanded §5.1 to discuss this limitation explicitly and to outline how the observed cross-family organization (transcending chemical similarity) already distinguishes the manifold from simple selection effects, while noting that future validation on newly synthesized materials is planned. revision: partial

Circularity Check

0 steps flagged

Unsupervised dimensionality reduction on ICSD descriptors yields post-hoc Tc correlation with no definitional reduction

full rationale

The derivation begins with unsupervised nonlinear dimensionality reduction applied to microscopic descriptors drawn from the ICSD (a structural database containing no Tc values). The resulting low-dimensional embedding is then observed to segregate superconductors and correlate with Tc. Because the reduction step itself is unsupervised and does not incorporate Tc or any fitted superconducting property, the subsequent correlation is an independent empirical finding rather than a quantity recovered by construction. No self-citation chain, ansatz smuggling, or renaming of known results is required to reach the reported organization; the central claim therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

Review performed on abstract only; full methods, equations, and validation unavailable, so ledger entries are provisional and minimal.

axioms (1)
  • domain assumption ICSD entries are sufficiently representative of crystalline materials for the discovered manifold to generalize
    Invoked implicitly when claiming the embedding organizes the materials landscape broadly.
invented entities (1)
  • emergent low-dimensional manifold of quantum materials no independent evidence
    purpose: To capture variance in material properties and govern Tc
    Introduced via the dimensionality reduction procedure; no independent falsifiable handle supplied in abstract.

pith-pipeline@v0.9.1-grok · 5731 in / 1264 out tokens · 19962 ms · 2026-06-27T07:50:17.884052+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

77 extracted references · 6 linked inside Pith

  1. [1]

    Triple product of atomic features, e.g., atomic mass of the first atom times covalent radius of the second atom times electronegativity of the third atom (165 features)

  2. [2]

    Two atomic features multiplied by the bond length connecting them, e.g., atomic mass of the first atom times covalent radius of the second atom times the length of the bond linking the two atoms (81 fea- tures)

  3. [3]

    Bond angle ( β), another geometric feature appear- ing for the first time in three-site clusters

  4. [4]

    Overall, there are 248 third-order features

    Bond angle β multiplied by the two bond lengths L1,L 2 forming it. Overall, there are 248 third-order features. For the three- dimensional embedding discussed in the main text, we include up to these third-order features. The resulting feature vector has 303 × (2 + nmoments) entries, and since we take nmoments = 5, our feature vector has 2 ,121 dimen- sio...

  5. [5]

    Jumper, R

    J. Jumper, R. Evans, A. Pritzel, T. Green, M. Figurnov, O. Ronneberger, K. Tunyasuvunakool, R. Bates, A. ˇZ´ ıdek, A. Potapenko, A. Bridgland, C. Meyer, S. A. A. Kohl, A. J. Ballard, A. Cowie, B. Romera-Paredes, S. Nikolov, R. Jain, J. Adler, T. Back, S. Petersen, D. Reiman, E. Clancy, M. Zielinski, M. Steinegger, M. Pacholska, T. Berghammer, S. Bodenstei...

    2021

  6. [6]

    Tunyasuvunakool, J

    K. Tunyasuvunakool, J. Adler, Z. Wu, T. Green, M. Zielin- ski, A. ˇZ´ ıdek, A. Bridgland, A. Cowie, C. Meyer, A. Lay- don, S. Velankar, G. J. Kleywegt, A. Bateman, R. Evans, A. Pritzel, M. Figurnov, O. Ronneberger, R. Bates, S. A. A. Kohl, A. Potapenko, A. J. Ballard, B. Romera- Paredes, S. Nikolov, R. Jain, E. Clancy, D. Reiman, S. Pe- tersen, A. W. Seni...

    2021

  7. [7]

    Noy and W

    S. Noy and W. Zhang, Science381, 187 (2023)

    2023

  8. [8]

    Vaswani, N

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. ukasz Kaiser, and I. Polosukhin, inAd- vances in Neural Information Processing Systems, Vol. 30 (Curran Associates, Inc., 2017)

    2017

  9. [9]

    OpenAI, GPT-4 Technical Report (2024), arXiv:2303.08774 [cs]

    Pith/arXiv arXiv 2024

  10. [10]

    K. Meng, D. Bau, A. Andonian, and Y. Belinkov, Ad- vances in Neural Information Processing Systems35, 17359 (2022)

    2022

  11. [11]

    Cunningham, A

    H. Cunningham, A. Ewart, L. Riggs, R. Huben, and L. Sharkey, Sparse Autoencoders Find Highly Interpretable Features in Language Models (2023), arXiv:2309.08600 [cs]

    Pith/arXiv arXiv 2023

  12. [12]

    R. Rao, J. Meier, T. Sercu, S. Ovchinnikov, and A. Rives, Transformer protein language models are unsupervised structure learners (2020)

    2020

  13. [13]

    Zhang, H

    Z. Zhang, H. K. Wayment-Steele, G. Brixi, H. Wang, D. Kern, and S. Ovchinnikov, Proceedings of the National Academy of Sciences121, e2406285121 (2024)

    2024

  14. [14]

    Zagorac, H

    D. Zagorac, H. M¨ uller, S. Ruehl, J. Zagorac, and S. Rehme, Journal of Applied Crystallography52, 918 (2019)

    2019

  15. [15]

    Sommer, R

    T. Sommer, R. Willa, J. Schmalian, and P. Friederich, Scientific Data10, 816 (2023)

    2023

  16. [16]

    P. Chen, L. Peng, R. Jiao, Q. Mo, Z. Wang, W. Huang, Y. Liu, and Y. Lu, Advances in Neural Information Pro- cessing Systems37, 108902 (2024)

    2024

  17. [17]

    Merchant, S

    A. Merchant, S. Batzner, S. S. Schoenholz, M. Aykol, G. Cheon, and E. D. Cubuk, Nature624, 80 (2023). 12

    2023

  18. [18]

    Gashmard, H

    H. Gashmard, H. Shakeripour, and M. Alaei, Scientific Reports14, 3965 (2024)

    2024

  19. [19]

    Cheng, C.-L

    M. Cheng, C.-L. Fu, R. Okabe, A. Chotrattanapituk, A. Boonkird, N. T. Hung, and M. Li, AI-driven materials design: A mini-review (2025), arXiv:2502.02905 [cond- mat]

    arXiv 2025

  20. [20]

    Cheng, W

    M. Cheng, W. Luo, H. Tang, B. Yu, Y. Cheng, W. Xie, J. Li, H. J. Kulik, and M. Li, Enhancing Materials Dis- covery with Valence Constrained Design in Generative Modeling (2025), arXiv:2507.19799 [cond-mat]

    arXiv 2025

  21. [21]

    F. Wang, A. G. Iwanicki, A. T. Sose, L. A. Pressley, T. M. McQueen, and S. A. Deshmukh, npj Computational Materials11, 124 (2025)

    2025

  22. [22]

    A. Ma, Y. Zhang, T. Christensen, H. C. Po, L. Jing, L. Fu, and M. Soljaˇ ci´ c, Nano Letters23, 772 (2023)

    2023

  23. [23]

    Ma and M

    A. Ma and M. Soljaˇ ci´ c, Learning simple heuristic rules for classifying materials based on chemical composition (2025), arXiv:2505.02361 [cond-mat]

    arXiv 2025

  24. [24]

    Lesser, Y

    O. Lesser, Y. Liu, N. Maus, A. Panigrahi, K. Mallayya, A. Gong, A. Kabra, S. B. Lee, S. Chatterjee, A. Merino, K. Q. Weinberger, L. M. Schoop, J. R. Gardner, and E.-A. Kim, Electron affinity difference distributions guide the discovery of the superconductor PtPb 3Bi (2026), arXiv:2510.07373 [cond-mat.supr-con]

    arXiv 2026

  25. [25]

    Panigrahi, Y

    A. Panigrahi, Y. Liu, O. Lesser, K. Mallayya, and E.-A. Kim, Graphlet Histogram Representation Database of In- organic Crystals (2026), arXiv:2606.10195 [cond-mat.mtrl- sci]

    Pith/arXiv arXiv 2026

  26. [26]

    Stanev, C

    V. Stanev, C. Oses, A. G. Kusne, E. Rodriguez, J. Paglione, S. Curtarolo, and I. Takeuchi, npj Com- putational Materials4, 1 (2018)

    2018

  27. [27]

    Hamidieh, Computational Materials Science154, 346 (2018)

    K. Hamidieh, Computational Materials Science154, 346 (2018)

    2018

  28. [28]

    Roter and S

    B. Roter and S. Dordevic, Physica C: Superconductivity and its Applications575, 1353689 (2020)

    2020

  29. [29]

    Pereti, K

    C. Pereti, K. Bernot, T. Guizouarn, F. Laufek, A. Vy- mazalov´ a, L. Bindi, R. Sessoli, and D. Fanelli, npj Com- putational Materials9, 71 (2023)

    2023

  30. [30]

    E. A. Pogue, A. New, K. McElroy, N. Q. Le, M. J. Pekala, I. McCue, E. Gienger, J. Domenico, E. Hedrick, T. M. Mc- Queen, B. Wilfong, C. D. Piatko, C. R. Ratto, A. Lennon, C. Chung, T. Montalbano, G. Bassen, and C. D. Stiles, npj Computational Materials9, 181 (2023)

    2023

  31. [31]

    Kaplan, A

    D. Kaplan, A. Zheng, J. Blawat, R. Jin, R. J. Cava, V. Oudovenko, G. Kotliar, A. M. Sengupta, and W. Xie, The European Physical Journal Plus140, 58 (2025)

    2025

  32. [32]

    S. R. Xie, Y. Quan, A. C. Hire, B. Deng, J. M. DeStefano, I. Salinas, U. S. Shah, L. Fanfarillo, J. Lim, J. Kim, G. R. Stewart, J. J. Hamlin, P. J. Hirschfeld, and R. G. Hennig, npj Computational Materials8, 14 (2022)

    2022

  33. [33]

    J. B. Gibson, A. C. Hire, P. M. Dee, B. Geisler, J. S. Kim, Z. Li, J. J. Hamlin, G. R. Stewart, P. J. Hirschfeld, and R. G. Hennig, Developing a Complete AI-Accelerated Workflow for Superconductor Discovery (2025), arXiv:2503.20005 [cond-mat]

    arXiv 2025

  34. [34]

    Prakash, J

    P. Prakash, J. B. Gibson, Z. Li, G. D. Gianluca, J. Es- quivel, E. Fuemmeler, B. Geisler, J. S. Kim, A. Roitberg, E. B. Tadmor, M. Liu, S. Martiniani, G. R. Stewart, J. J. Hamlin, P. J. Hirschfeld, and R. G. Hennig, Guided Dif- fusion for the Discovery of New Superconductors (2025), arXiv:2509.25186 [cond-mat]

    arXiv 2025

  35. [35]

    Tahmassebpur, S

    J. Tahmassebpur, S. Chaudhari, C. M´ endez, R. Choud- hary, S. Kundu, R. E. Schaak, H. Abru˜ na, P. Frazier, and T. Arias, A seven-facet polyhedron captures the composition-only formation-energy landscape of inorganic solids (2026), arXiv:2602.00254 [cond-mat]

    arXiv 2026

  36. [36]

    McInnes, J

    L. McInnes, J. Healy, and J. Melville, arXiv preprint arXiv:1802.03426 (2018)

    Pith/arXiv arXiv 2018

  37. [37]

    W. Ye, X. Lei, M. Aykol, and J. H. Montoya, Scientific Data9, 302 (2022)

    2022

  38. [38]

    B. D. Lee, J.-W. Lee, W. B. Park, J. Park, M.-Y. Cho, S. Pal Singh, M. Pyo, and K.-S. Sohn, Advanced Intelli- gent Systems4, 2200042 (2022)

    2022

  39. [39]

    Arvanitidis, L

    G. Arvanitidis, L. K. Hansen, and S. Hauberg, arXiv preprint arXiv:1710.11379 (2017)

    arXiv 2017

  40. [40]

    J. Z. Kim, N. Perrin-Gilbert, E. Narmanli, P. Klein, C. R. Myers, I. Cohen, J. J. Waterfall, and J. P. Sethna, arXiv preprint arXiv:2403.01078 (2024)

    arXiv 2024

  41. [41]

    Scarselli, M

    F. Scarselli, M. Gori, A. C. Tsoi, M. Hagenbuchner, and G. Monfardini, IEEE Transactions on Neural Networks 20, 61 (2009)

    2009

  42. [42]

    Isayev, C

    O. Isayev, C. Oses, C. Toher, E. Gossett, S. Curtarolo, and A. Tropsha, Nature Communications8, 15679 (2017)

    2017

  43. [43]

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

    Pith/arXiv arXiv 2021

  44. [44]

    Reiser, M

    P. Reiser, M. Neubert, A. Eberhard, L. Torresi, C. Zhou, C. Shao, H. Metni, C. van Hoesel, H. Schopmans, T. Som- mer, and P. Friederich, Communications Materials3, 93 (2022)

    2022

  45. [45]

    V. Fung, J. Zhang, E. Juarez, and B. G. Sumpter, npj Computational Materials7, 84 (2021)

    2021

  46. [46]

    Zhang, Z

    X. Zhang, Z. Zhou, C. Ming, and Y.-Y. Sun, The Journal of Physical Chemistry Letters14, 11342 (2023)

    2023

  47. [47]

    Choudhary and B

    K. Choudhary and B. DeCost, npj Computational Mate- rials7, 185 (2021)

    2021

  48. [48]

    D. P. Kingma, M. Welling,et al., Foundations and Trends®in Machine Learning12, 307 (2019)

    2019

  49. [49]

    P. Li, Y. Pei, and J. Li, Applied Soft Computing138, 110176 (2023)

    2023

  50. [50]

    Hornik, M

    K. Hornik, M. Stinchcombe, and H. White, Neural net- works2, 359 (1989)

    1989

  51. [51]

    MDR: NIMS Materials Data Repository, https://mdr.nims.go.jp

  52. [52]

    Pearson, Proceedings of the Royal Society of London 58, 240 (1997)

    K. Pearson, Proceedings of the Royal Society of London 58, 240 (1997)

    1997

  53. [53]

    I. T. Jolliffe and J. Cadima, Philosophical Transactions of the Royal Society A: Mathematical, Physical and En- gineering Sciences374, 20150202 (2016)

    2016

  54. [54]

    R. J. Freund and W. J. Wilson,Statistical Methods(Else- vier, 2003)

    2003

  55. [55]

    C. E. Rasmussen and C. K. I. Williams,Gaussian Pro- cesses for Machine Learning, Adaptive Computation and Machine Learning (MIT Press, Cambridge, Mass, 2006)

    2006

  56. [56]

    Kohn and L

    W. Kohn and L. J. Sham, Physical Review140, A1133 (1965)

    1965

  57. [57]

    R. G. Parr and Y. Weitao,Density-Functional Theory of Atoms and Molecules (International Series of Monographs on Chemistry)(Oxford University Press, USA, 1994)

    1994

  58. [58]

    Marsiglio, Annals of Physics Eliashberg Theory at 60: Strong-coupling Superconductivity and Beyond,417, 168102 (2020)

    F. Marsiglio, Annals of Physics Eliashberg Theory at 60: Strong-coupling Superconductivity and Beyond,417, 168102 (2020)

    2020

  59. [59]

    Sanna, C

    A. Sanna, C. Pellegrini, and E. K. U. Gross, Physical Review Letters125, 057001 (2020)

    2020

  60. [60]

    Bharathi, S

    A. Bharathi, S. Jemima Balaselvi, M. Premila, T. N. Sairam, G. L. N. Reddy, C. S. Sundar, and Y. Hariharan, Solid State Communications124, 423 (2002)

    2002

  61. [61]

    A. M. Fogg, J. B. Claridge, G. R. Darling, and M. J. 13 Rosseinsky, Chemical Communications , 1348 (2003)

    2003

  62. [62]

    Rosner, A

    H. Rosner, A. Kitaigorodsky, and W. E. Pickett, Physical Review Letters88, 127001 (2002)

    2002

  63. [63]

    A. M. Fogg, J. Meldrum, G. R. Darling, J. B. Claridge, and M. J. Rosseinsky, Journal of the American Chemical Society128, 10043 (2006)

    2006

  64. [64]

    Akashi, K

    R. Akashi, K. Nakamura, R. Arita, and M. Imada, Physi- cal Review B86, 054513 (2012)

    2012

  65. [65]

    Borinaga, U

    M. Borinaga, U. Aseginolaza, I. Errea, M. Calandra, F. Mauri, and A. Bergara, Physical Review B96, 184505 (2017)

    2017

  66. [66]

    K. A. Szewczyk, I. A. Domagalska, A. P. Durajski, and R. Szczesniak, Beilstein Journal of Nanotechnology11, 1178 (2020)

    2020

  67. [67]

    Barroso-Luque, M

    L. Barroso-Luque, M. Shuaibi, X. Fu, B. M. Wood, M. Dzamba, M. Gao, A. Rizvi, C. L. Zitnick, and Z. W. Ulissi, Open Materials 2024 (OMat24) Inorganic Materials Dataset and Models (2024), arXiv:2410.12771 [cond-mat]

    Pith/arXiv arXiv 2024

  68. [68]

    Itani, Y

    S. Itani, Y. Zhang, and J. Zang, The Northeast Materials Database for Magnetic Materials (2025), arXiv:2409.15675 [cond-mat]

    arXiv 2025

  69. [69]

    Itani, Y

    S. Itani, Y. Zhang, and J. Zang, Large Language Model- Driven Database for Thermoelectric Materials (2024), arXiv:2501.00564 [cond-mat]

    arXiv 2024

  70. [70]

    Petralanda, Y

    U. Petralanda, Y. Jiang, B. A. Bernevig, N. Regnault, and L. Elcoro, Two-dimensional Topological Quantum Chemistry and Catalog of Topological Materials (2024), arXiv:2411.08950 [cond-mat]

    arXiv 2024

  71. [71]

    E. Y. Andrei, D. K. Efetov, P. Jarillo-Herrero, A. H. MacDonald, K. F. Mak, T. Senthil, E. Tutuc, A. Yazdani, and A. F. Young, Nature Reviews Materials6, 201 (2021)

    2021

  72. [72]

    K. P. Nuckolls and A. Yazdani, Nature Reviews Materials 9, 460 (2024)

    2024

  73. [73]

    H. Park, A. Onwuli, K. T. Butler, and A. Walsh, Faraday Discussions256, 601 (2025)

    2025

  74. [74]

    Van der Maaten and G

    L. Van der Maaten and G. Hinton, Journal of Machine Learning Research9(2008)

    2008

  75. [75]

    Kobak and G

    D. Kobak and G. C. Linderman, Nature Biotechnology 39, 156 (2021)

    2021

  76. [76]

    Meil˘ a and H

    M. Meil˘ a and H. Zhang, Annual Review of Statistics and Its Application11, 393 (2024)

    2024

  77. [77]

    Choquet-Bruhat, C

    Y. Choquet-Bruhat, C. DeWitt-Morette, and M. Dillard- Bleick,Analysis, Vol. 1 (Gulf Professional Publishing, 1982). 14 Embedding Iron- Cuprate Carbon Chevrel Heavy Other based Fermion SC 3D embedding, chemical composition features−10.01−43.41−11.05−1.38−15.88−48.77 2D embedding, third-order graph features−10.09−19.27−5.86−3.84−9.13−13.25 3D embedding, fir...

    1982