Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Hybrid Physics-Machine Learning Models for Quantitative Electron Diffraction Refinements

T0 review · 3 major / 4 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read This paper claims that a differentiable hybrid of physics simulation and neural networks can jointly refine physical parameters and experimental variables in 3D electron diffraction, achieving state-of-the-art recovery of atomic positions,

desk verdict Promising hybrid differentiable framework for 3D-ED refinement, but the abstract alone can't back the SOTA claims and the identifiability of the NN corrections is the key thing to check. read the letter →

arxiv 2508.05908 v1 pith:EDZVKYVS submitted 2025-08-08 physics.comp-ph cs.LG

Shreshth A. Malik , Tiarnan A.S. Doherty , Benjamin Colmey , Stephen J. Roberts , Yarin Gal , Paul A. Midgley This is my paper · ORCID
classification physics.comp-phcs.LG
keywords electrondiffraction3D-EDstructurerefinementhybridphysics-machinelearningdifferentiablesimulationautomaticdifferentiationneuralnetworksthicknessdistribution
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 tries to establish that quantitative electron diffraction refinements no longer require analytic models for every experimental effect. The authors combine a differentiable physics simulation with neural networks that represent experimental variables, and optimize both together using automatic differentiation. Applied to three-dimensional electron diffraction, the method learns complex thickness distributions directly from data and recovers atomic positions and thermal displacements with state-of-the-art accuracy on synthetic and experimental datasets. If correct, this makes structure refinement more scalable than traditional second-order methods and opens the same hybrid strategy to other electron microscopy techniques.

What carries the argument

The key mechanism is an end-to-end differentiable simulation pipeline for 3D electron diffraction in which neural-network modules parameterize experimental variables such as thickness distributions. Automatic differentiation routes gradients from the measured intensities back through the physics model into both the physical parameters (atomic positions, thermal displacements) and the network weights, enabling joint first-order optimization. The design deliberately assigns the neural network the role of experimental correction, so the physical model does not need to be analytically corrected for geometry.

What would settle it

Perform the refinement on a simulated crystal with known atomic positions while systematically varying the assumed thickness model; if the neural network absorbs structural signal, the recovered positions will drift systematically as the thickness prior changes. An experimental check would compare the network-learned thickness distribution to an independently measured thickness profile (e.g., from electron energy-loss spectroscopy) on the same specimen: disagreement correlated with structural refinement error would indicate a trade-off between the network and the physics.

Watch

Extended reading notes

Core claim

The central claim is that a hybrid physics-machine learning framework can solve a long-standing limitation of quantitative electron diffraction: experimental effects that are hard to model analytically, such as thickness distributions, can be absorbed by neural-network components and learned directly from diffraction intensities. By making the entire simulation pipeline differentiable, the paper shows that gradient-based joint optimization of physical parameters and network weights is possible, and demonstrates that this recovers atomic positions, thermal displacement parameters, and thickness profiles with high fidelity. This positions differentiable hybrid modeling as a new paradigm for qu

Load-bearing premise

The load-bearing premise is that the neural-network component can be constrained to represent experimental effects without absorbing signal from the physical parameters being refined; if the two are not identifiable from the diffraction intensities, the jointly fitted atomic positions and displacements would be biased.

Editorial extensions

If this is right

  • The method recovers atomic positions, thermal displacements, and thickness profiles from 3D-ED data with state-of-the-art accuracy on synthetic and experimental datasets.
  • The architecture is modular and can be extended to additional physical phenomena and other electron microscopy techniques.
  • The differentiable, first-order approach scales more favorably than second-order optimization methods traditionally used in refinement.
  • Learned thickness distributions remove the need for simplified geometric models, eliminating a common source of systematic error in quantitative diffraction analysis.

Reading between the lines

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

  • A natural but untested extension is to apply the same hybrid strategy to X-ray or neutron diffraction, where absorption and extinction corrections are similarly hard to model analytically.
  • The identifiability of the network versus the physical model is the key open question; a transfer test (keeping the network fixed when the physics changes) would probe whether the network truly captures experimental effects or absorbs structural signal.
  • The framework implies that inelastic scattering, beam misalignment, or detector nonlinearities could also be represented as learned experimental variables, but the paper does not demonstrate those cases.
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. The manuscript proposes a hybrid physics-machine-learning framework for quantitative three-dimensional electron diffraction (3D-ED) refinement. A differentiable physical simulation is coupled to a neural-network component intended to represent experimental effects, and atomic positions, thermal displacement parameters, and thickness distributions are recovered by gradient-based joint optimization. The abstract claims state-of-the-art refinement performance on synthetic and experimental data, superior scalability relative to traditional second-order methods, and modular extensibility to other electron microscopy techniques. The supplied main text is an undecodable/corrupted encoding, so the assessment below rests on the abstract and on the reader's report.

Significance. If substantiated, the proposed framework would be a meaningful contribution: end-to-end differentiable refinement with automatic differentiation could avoid the computational cost of second-order optimizers, and learning thickness distributions from data rather than assuming simplified geometric models is a useful capability. The modular architecture is a plausible template for other quantitative EM problems. Credit is due for framing the core idea clearly and for making the problem of experimental corrections a first-class modeling target. However, the version available for review provides no numerical evidence, no identifiability analysis, and no scalability measurements, so the central claims remain unverified. The reader's concern about the neural network absorbing structural signal is the most load-bearing risk; it is not addressed in the abstract.

major comments (3)
  1. [Abstract] The central claim—that joint optimization recovers unbiased atomic positions and thermal displacements—requires an identifiability analysis. The Debye-Waller factor exp(-2B s^2) produces a smooth modulation in scattering-vector space, and thickness-dependent dynamical effects can produce similar smooth intensity variations. Without explicit constraints, regularization, or a demonstration of orthogonality, a flexible neural network can absorb part of the structural signal, biasing the refined parameters. The authors should provide synthetic ground-truth tests where the true structural parameters are known and show that the NN correction does not trade off against changes in B or atomic positions, e.g., by varying NN capacity and measuring parameter bias.
  2. [Abstract] The claim 'state-of-the-art refinement performance across synthetic and experimental datasets' is unsupported by any numerical results in the accessible text. There are no reported error bars, no comparison baselines, no dataset sizes, and no metrics. Because this claim is the paper's headline, the authors must include a quantitative comparison with existing refinement pipelines, including accuracy and precision for positions and displacement parameters, on both synthetic and experimental data.
  3. [Abstract] The assertion of 'superior scalability compared to traditional second-order methods' is not backed by any complexity analysis or empirical timing/memory comparison. Gradient-based first-order methods have lower per-iteration cost but may require many more iterations; the total cost and robustness should be measured on a representative problem. Without this, the scalability claim remains an assertion.
minor comments (4)
  1. [Full text] The main text as supplied is not readable—it appears as a corrupted/undecodable character encoding. A properly rendered PDF or source text must be provided before the paper can be evaluated in detail.
  2. [Abstract] The abstract uses 'state-of-the-art' and 'high fidelity' without defining the comparison standard or the accuracy metric. Please quantify these terms or qualify them with specific numerical results.
  3. [Abstract] The phrase 'learns complex thickness distributions directly from diffraction data' should be clarified: how is the thickness distribution parameterized (discrete layers, continuous profile, neural field)? This affects the identifiability discussion and should be stated explicitly.
  4. [General] The acronym '3D-ED' is used in the abstract; it should be expanded at first use. Also, the link to 'other electron microscopy techniques' would be strengthened by naming one or two concrete extensions in the abstract or introduction.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identified: the hybrid refinement is a joint fit against data, not a derivation that presupposes its conclusion.

full rationale

The abstract describes a hybrid physics+ML refinement framework in which differentiable simulations and neural-network components are jointly optimized against diffraction data. The recovered quantities (atomic positions, thermal displacements, thickness profiles) are outputs of this optimization, not inputs smuggled in as predictions. The neural network is explicitly described as representing experimental variables and as learning thickness distributions from data; this is straightforward fitting/estimation, not a disguised restatement of the target. No equation or passage in the available text defines a physical parameter in terms of the quantity it is supposed to predict, and no load-bearing self-citation or imported uniqueness theorem appears. The identifiability concern (that a flexible NN could absorb structural signal) is a genuine correctness risk, but it is not circularity: it does not make the claimed recovery true by construction. With only the abstract available and the body text corrupted, no specific circular reduction can be quoted, and per the hard rules circularity should not be inferred without such evidence.

Assumptions & free parameters 1 free parameters · 3 assumptions · 1 invented entities

Since only the abstract is available, the ledger is inferred from the stated design. The main unverified assumptions concern the accuracy of the physical forward model and the identifiability of the neural-network corrections.

free parameters (1)
  • Neural network weights and biases = Unknown, trained on data
    The ML component representing experimental effects is fitted jointly with physical parameters to the diffraction data. The central claim depends on this fitted component not corrupting the structural refinement.
assumptions (3)
  • domain assumption The forward physical simulation accurately models electron scattering (kinematical or dynamical diffraction theory).
    The refinement optimizes physical parameters inside this simulation; if the simulation's approximations are wrong, recovered structures will be biased. Invoked implicitly by the whole framework.
  • ad hoc to paper Experimental effects not captured by the physical simulation can be represented by a neural network trained on the same data.
    This is the key modeling bet of the paper. The network is flexible enough to absorb unmodeled physics, and the optimization separates it from the structural parameters.
  • domain assumption Automatic differentiation through the simulation provides correct gradients for joint optimization.
    The claimed scalability and convergence rely on the correctness and stability of AD through the forward model.
invented entities (1)
  • Neural network correction term for experimental effects
    purpose: Absorbs residual mismatches between the physical simulation and real-world diffraction data, e.g., sample thickness distributions.
    The NN is fitted to the same data used for structure refinement, so it has no independent falsifiable handle. It could in principle absorb structural signal, making the physical parameters less trustworthy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid Physics-Machine Learning Models for Quantitative Electron Diffraction Refinements." pith.science (2026). https://pith.science/paper/EDZVKYVS

@misc{pith2026250805908,
  author       = {Pith},
  title        = {Pith review of: Hybrid Physics-Machine Learning Models for Quantitative Electron Diffraction Refinements},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EDZVKYVS}},
  note         = {Machine review of arXiv:2508.05908}
}
read the original abstract

High-fidelity electron microscopy simulations required for quantitative crystal structure refinements face a fundamental challenge: while physical interactions are well-described theoretically, real-world experimental effects are challenging to model analytically. To address this gap, we present a novel hybrid physics-machine learning framework that integrates differentiable physical simulations with neural networks. By leveraging automatic differentiation throughout the simulation pipeline, our method enables gradient-based joint optimization of physical parameters and neural network components representing experimental variables, offering superior scalability compared to traditional second-order methods. We demonstrate this framework through application to three-dimensional electron diffraction (3D-ED) structure refinement, where our approach learns complex thickness distributions directly from diffraction data rather than relying on simplified geometric models. This method achieves state-of-the-art refinement performance across synthetic and experimental datasets, recovering atomic positions, thermal displacements, and thickness profiles with high fidelity. The modular architecture proposed can naturally be extended to accommodate additional physical phenomena and extended to other electron microscopy techniques. This establishes differentiable hybrid modeling as a powerful new paradigm for quantitative electron microscopy, where experimental complexities have historically limited analysis.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. The role of absorption in three-dimensional electron diffraction dynamical structure refinement

    cond-mat.mtrl-sci 2026-02 conditional novelty 6.0 of 10

    Absorption in 3D electron diffraction refinement is a uniform dimming that can be ignored in routine work, except for high-Z materials at thicknesses approaching the extinction distance, where reflection-specific 'ano...

Reference graph

Works this paper leans on

56 extracted references · 27 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline " cite write " FUNCTION editor.postfix editor num.names #1 > "( )" "( )" if FUNCTION editor.trans.postfix editor num.names #1 > "( )" "( )" if FUNCTION trans.postfix translator num.names #1 > "( )" "( )" if FUNCTION authors.editors.reflist.apa5 'field := 'dot := field num.names 'numnames := numnames 'format.num.names := format.num.names na...

  2. [2]

    sn-aps.bst

    FUNCTION identify.aps.version "sn-aps.bst" " [2024/07/19 v1.1 APS bibliography style]" * top ENTRY address author booktitle chapter doi edition editor eid howpublished institution journal key keywords month note number organization pages publisher school series title type url volume year eprint archive archivePrefix primaryClass adsurl adsnote version lab...

    2024

  3. [3]

    write newline

    " write newline "" before.all 'output.state := FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap duplicate "7" = swap duplicate "8" = swap "9" = or or or or or or or or or FUNCTION n.separate 't := "" #0 'numnames := t empty not t #-1 #1 subs...

  4. [4]

    , Bengio , Y

    bbook Goodfellow , I. , Bengio , Y. , Courville , A. : Deep Learning . MIT Press , Cambridge, MA ( 2016 ). http://www.deeplearningbook.org/ bbook

    work page 2016

  5. [5]

    , Sutskever , I

    bchapter Krizhevsky , A. , Sutskever , I. , Hinton , G.E. : Imagenet classification with deep convolutional neural networks . In: Bartlett , P.L. , Pereira , F.C.N. , Burges , C.J.C. , Bottou , L. , Weinberger , K.Q. (eds.) Advances in Neural Information Processing Systems 25: 26th Annual Conference on Neural Information Processing Systems 2012. Proceedin...

    work page 2012

  6. [6]

    , Zhang , X

    bchapter He , K. , Zhang , X. , Ren , S. , Sun , J. : Deep residual learning for image recognition . In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pp. 770 -- 778 ( 2016 ) bchapter

    work page 2016

  7. [7]

    , Hubert , T

    botherref Silver , D. , Hubert , T. , Schrittwieser , J. , Antonoglou , I. , Lai , M. , Guez , A. , Lanctot , M. , Sifre , L. , Kumaran , D. , Graepel , T. , et al.: Mastering chess and shogi by self-play with a general reinforcement learning algorithm. arXiv preprint arXiv:1712.01815 (2017) botherref

    arXiv 2017

  8. [8]

    , Antonoglou , I

    barticle Schrittwieser , J. , Antonoglou , I. , Hubert , T. , Simonyan , K. , Sifre , L. , Schmitt , S. , Guez , A. , Lockhart , E. , Hassabis , D. , Graepel , T. , Lillicrap , T.P. , Silver , D. : Mastering A tari, G o, chess and shogi by planning with a learned model . Nat. 588 ( 7839 ), 604 -- 609 ( 2020 ) 10.1038/S41586-020-03051-4 barticle

    doi:10.1038/s41586-020-03051-4 2020

Show all 56 references
  1. [9]

    , Stinchcombe , M

    barticle Hornik , K. , Stinchcombe , M. , White , H. : Multilayer feedforward networks are universal approximators . Neural networks 2 ( 5 ), 359 -- 366 ( 1989 ) barticle

    1989

  2. [10]

    , Smith , K

    botherref Avila Belbute-Peres , F. , Smith , K. , Allen , K. , Tenenbaum , J. , Kolter , J.Z. : End-to-end differentiable physics for learning and control. Advances in neural information processing systems 31 (2018) botherref

    2018

  3. [11]

    , Lopez-Guevara , T

    botherref Allen , K.R. , Lopez-Guevara , T. , Stachenfeld , K. , Sanchez-Gonzalez , A. , Battaglia , P. , Hamrick , J. , Pfaff , T. : Physical design using differentiable learned simulators. arXiv preprint arXiv:2202.00728 (2022) botherref

    2022 arXiv

  4. [12]

    , Fortunato , M

    botherref Pfaff , T. , Fortunato , M. , Sanchez-Gonzalez , A. , Battaglia , P.W. : Learning mesh-based simulation with graph networks. arXiv preprint arXiv:2010.03409 (2020) botherref

    2010 arXiv

  5. [13]

    , Szunyogh , I

    barticle Arcomano , T. , Szunyogh , I. , Wikner , A. , Pathak , J. , Hunt , B.R. , Ott , E. : A hybrid approach to atmospheric modeling that combines machine learning with a physics-based numerical model . Journal of Advances in Modeling Earth Systems 14 ( 3 ), 2021 -- 002712 ...

    2021

  6. [14]

    , Caro , M.A

    barticle Deringer , V.L. , Caro , M.A. , Cs \'a nyi , G. : Machine learning interatomic potentials as emerging tools for materials science . Advanced Materials 31 ( 46 ), 1902765 ( 2019 ) barticle

    2019

  7. [15]

    , Cubuk , E.D

    barticle Schoenholz , S. , Cubuk , E.D. : JAX MD : a framework for differentiable physics . Advances in Neural Information Processing Systems 33 , 11428 -- 11441 ( 2020 ) barticle

    2020

  8. [16]

    , Pearlmutter , B.A

    barticle Baydin , A.G. , Pearlmutter , B.A. , Radul , A.A. , Siskind , J.M. : Automatic differentiation in machine learning: a survey . Journal of Marchine Learning Research 18 , 1 -- 43 ( 2018 ) barticle

    2018

  9. [17]

    , Broek , W

    barticle Pennington , R.S. , Broek , W. , Koch , C.T. : Third-dimension information retrieval from a single convergent-beam transmission electron diffraction pattern using an artificial neural network . Physical Review B 89 ( 20 ), 205409 ( 2014 ) 10.1103/PhysRevB.89.205409 . ...

    2014 doi

  10. [18]

    , Koch , C.T

    barticle Van Den Broek , W. , Koch , C.T. : Method for Retrieval of the Three - Dimensional Object Potential by Inversion of Dynamical Electron Scattering . Physical Review Letters 109 ( 24 ), 245502 ( 2012 ) 10.1103/PhysRevLett.109.245502 . Accessed 2024-01-12 barticle

    2012 doi

  11. [19]

    , Herdegen , Z

    barticle Diederichs , B. , Herdegen , Z. , Strauch , A. , Filbir , F. , Müller-Caspary , K. : Exact inversion of partially coherent dynamical electron scattering for picometric structure retrieval . Nature Communications 15 ( 1 ), 101 ( 2024 ) 10.1038/s41467-023-44268-x . Publ...

    2024 doi

  12. [20]

    : The scattering of fast electrons by crystals

    barticle Humphreys , C.J. : The scattering of fast electrons by crystals . Reports on Progress in Physics 42 ( 11 ), 1825 ( 1979 ) 10.1088/0034-4885/42/11/002 . Accessed 2023-09-28 barticle

    1979 doi

  13. [21]

    : Advanced Computing in Electron Microscopy

    bbook Kirkland , E.J. : Advanced Computing in Electron Microscopy . Springer , Boston, MA ( 2010 ). 10.1007/978-1-4419-6533-2 . https://link.springer.com/10.1007/978-1-4419-6533-2 Accessed 2024-01-11 bbook

    2010 doi

  14. [22]

    , Brázda , P

    barticle Palatinus , L. , Brázda , P. , Boullay , P. , Perez , O. , Klementová , M. , Petit , S. , Eigner , V. , Zaarour , M. , Mintova , S. : Hydrogen positions in single nanocrystals revealed by electron diffraction . Science 355 ( 6321 ), 166 -- 169 ( 2017 ) 10.1126/science...

    2017 doi

  15. [23]

    : The potential and limitations of neutrons, electrons and x-rays for atomic resolution microscopy of unstained biological molecules

    barticle Henderson , R. : The potential and limitations of neutrons, electrons and x-rays for atomic resolution microscopy of unstained biological molecules . Quarterly reviews of biophysics 28 ( 2 ), 171 -- 193 ( 1995 ) barticle

    1995

  16. [24]

    , Shi , D

    barticle Nannenga , B.L. , Shi , D. , Leslie , A.G. , Gonen , T. : High-resolution structure determination by continuous-rotation data collection in microed . Nature methods 11 ( 9 ), 927 -- 930 ( 2014 ) barticle

    2014

  17. [25]

    , Moodie , A.F

    barticle Cowley , J.M. , Moodie , A.F. : The scattering of electrons by atoms and crystals. I . A new theoretical approach . Acta Crystallographica 10 ( 10 ), 609 -- 619 ( 1957 ) 10.1107/S0365110X57002194 barticle

    1957 doi

  18. [26]

    , Palatinus , L

    barticle Brázda , P. , Palatinus , L. , Babor , M. : Electron diffraction determines molecular absolute configuration in a pharmaceutical nanocrystal . Science 364 ( 6441 ), 667 -- 669 ( 2019 ) 10.1126/science.aaw2560 . Publisher: American Association for the Advancement of Sc...

    2019 doi

  19. [27]

    , Mugnaioli , E

    barticle Gemmi , M. , Mugnaioli , E. , Gorelik , T.E. , Kolb , U. , Palatinus , L. , Boullay , P. , Hovmöller , S. , Abrahams , J.P. : 3D Electron Diffraction : The Nanocrystallography Revolution . ACS Central Science 5 ( 8 ), 1315 -- 1329 ( 2019 ) 10.1021/acscentsci.9b00394 ....

    2019 doi

  20. [28]

    , Yörük , E

    barticle Suresh , A. , Yörük , E. , Cabaj , M.K. , Brázda , P. , Výborný , K. , Sedláček , O. , Müller , C. , Chintakindi , H. , Eigner , V. , Palatinus , L. : Ionisation of atoms determined by kappa refinement against 3D electron diffraction data . Nature Communications 15 ( ...

    2024 doi

  21. [29]

    : Zur theorie des durchgangs schneller korpuskularstrahlen durch materie

    barticle Bethe , H. : Zur theorie des durchgangs schneller korpuskularstrahlen durch materie . Annalen der Physik 392 , 55 -- 129 ( 1928 ) 10.1002/andp.19283920202 barticle

    1928 doi

  22. [30]

    , Petříček , V

    barticle Palatinus , L. , Petříček , V. , Corrêa , C.A. : Structure refinement using precession electron diffraction tomography and dynamical diffraction: theory and implementation . Acta Crystallographica Section A: Foundations and Advances 71 ( 2 ), 235 -- 244 ( 2015 ) 10.11...

    2015 doi

  23. [31]

    , Spence , J.C.H

    barticle Zuo , J.M. , Spence , J.C.H. : Automated structure factor refinement from convergent-beam patterns . Ultramicroscopy 35 ( 3 ), 185 -- 196 ( 1991 ) 10.1016/0304-3991(91)90071-D . Accessed 2025-05-01 barticle

    1991 doi

  24. [32]

    , Hinton , G.E

    barticle Rumelhart , D.E. , Hinton , G.E. , Williams , R.J. : Learning representations by back-propagating errors . Nature 323 ( 6088 ), 533 -- 536 ( 1986 ) 10.1038/323533a0 . Publisher: Nature Publishing Group . Accessed 2025-04-30 barticle

    1986 doi

  25. [33]

    , Brázda , P

    barticle Palatinus , L. , Brázda , P. , Jelínek , M. , Hrdá , J. , Steciuk , G. , Klementová , M. : Specifics of the data processing of precession electron diffraction tomography data and their implementation in the program PETS2 .0 . Acta Crystallographica Section B: Structur...

    2019 doi

  26. [34]

    , Jacob , D

    barticle Palatinus , L. , Jacob , D. , Cuvillier , P. , Klementová , M. , Sinkler , W. , Marks , L.D. : Structure refinement from precession electron diffraction data . Acta Crystallographica Section A: Foundations of Crystallography 69 ( 2 ), 171 -- 188 ( 2013 ) 10.1107/S0108...

    2013 doi

  27. [35]

    , Krysiak , Y

    barticle Klar , P.B. , Krysiak , Y. , Xu , H. , Steciuk , G. , Cho , J. , Zou , X. , Palatinus , L. : Accurate structure models and absolute configuration determination using dynamical effects in continuous-rotation 3D electron diffraction data . Nature Chemistry 15 ( 6 ), 848...

    2023 doi

  28. [36]

    , Ba , J

    bchapter Kingma , D.P. , Ba , J. : Adam: A method for stochastic optimization . In: Bengio , Y. , LeCun , Y. (eds.) 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings ( 2015 ). http://arxiv.org/a...

    2015 arXiv

  29. [37]

    http://pets.fzu.cz/ Accessed 2025-05-06 botherref

    botherref PETS . http://pets.fzu.cz/ Accessed 2025-05-06 botherref

    2025

  30. [38]

    : Modelling dynamical 3D electron diffraction intensities

    botherref Mendis , B. : Modelling dynamical 3D electron diffraction intensities. II . The role of inelastic scattering. Acta Crystallographica Section A 80(2) (2024) 10.1107/s2053273323010690 . Publisher: International Union of Crystallography. Accessed 2024-08-27 botherref

    2024 doi

  31. [39]

    Phase Scrambling

    botherref Mendis , B.G. : A “ Phase Scrambling ” Algorithm for Parallel Multislice Simulation of Multiple Phonon and Plasmon Scattering Configurations . Microscopy and Microanalysis 29(3) (2023) 10.1093/micmic/ozad052 . Publisher: Cambridge University Press. Accessed 2024-08-2...

    2023 doi

  32. [40]

    , Gross , S

    bchapter Paszke , A. , Gross , S. , Massa , F. , Lerer , A. , Bradbury , J. , Chanan , G. , Killeen , T. , Lin , Z. , Gimelshein , N. , Antiga , L. , Desmaison , A. , K \" o pf , A. , Yang , E.Z. , DeVito , Z. , Raison , M. , Tejani , A. , Chilamkurthy , S. , Steiner , B. , Fa...

    2019

  33. [41]

    , Howie , A

    bbook Hirsch , P.B. , Howie , A. , Nicholson , R.B. , Pashley , D.W. , Whelan , M.J. : Electron Microscopy of Thin Crystals , p. 549 . Butterworths , London ( 1965 ) bbook

    1965

  34. [42]

    , Zuo , J.-M

    bbook Spence , J.M. , Zuo , J.-M. : Electron Microdiffraction . Springer , New York ( 2020 ). 10.1007/978-1-4899-2353-0 bbook

    2020 doi

  35. [43]

    , Adams , P.D

    barticle Grosse-Kunstleve , R.W. , Adams , P.D. : On the handling of atomic anisotropic displacement parameters . Journal of Applied Crystallography 35 ( 4 ), 477 -- 480 ( 2002 ) 10.1107/S0021889802008580 barticle

    2002 doi

  36. [44]

    , Welling , M

    bchapter Kingma , D.P. , Welling , M. : Auto-encoding variational bayes . In: Bengio , Y. , LeCun , Y. (eds.) 2nd International Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14-16, 2014, Conference Track Proceedings ( 2014 ). http://arxiv.org/abs/...

    2014 arXiv

  37. [45]

    , Gommers , R

    barticle Virtanen , P. , Gommers , R. , Oliphant , T.E. , Haberland , M. , Reddy , T. , Cournapeau , D. , Burovski , E. , Peterson , P. , Weckesser , W. , Bright , J. , van der Walt , S.J. , Brett , M. , Wilson , J. , Millman , K.J. , Mayorov , N. , Nelson , A.R.J. , Jones , E...

    2020

  38. [46]

    sn-basic.bst

    FUNCTION identify.basic.version "sn-basic.bst" " [2024/07/19 v1.1 bibliography style]" * top ENTRY address archive author booktitle chapter doi edition editor eid eprint howpublished institution journal key keywords month note number organization pages publisher school series ...

    2024

  39. [47]

    write newline

    " write newline "" before.all 'output.state := FUNCTION add.period duplicate empty 'skip "." * add.blank if FUNCTION if.digit duplicate "0" = swap duplicate "1" = swap duplicate "2" = swap duplicate "3" = swap duplicate "4" = swap duplicate "5" = swap duplicate "6" = swap dupl...

  40. [48]

    write newline

    " write newline "" before.all 'output.state := FUNCTION output.doi doi empty skip "doi:" doi * "" * output if FUNCTION format.archive archivePrefix empty "" archivePrefix ":" * if FUNCTION format.primaryClass primaryClass empty "" " [" primaryClass * "] " * if FUNCTION format....

  41. [49]

    write newline

    " write newline "" before.all 'output.state := FUNCTION string.to.integer 't := t text.length 'k := #1 'char.num := t char.num #1 substring 's := s is.num s "." = or char.num k = not and char.num #1 + 'char.num := while char.num #1 - 'char.num := t #1 char.num substring FUNCTI...

  42. [50]

    write newline

    " write newline "" before.all 'output.state := FUNCTION string.to.integer 't := t text.length 'k := #1 'char.num := t char.num #1 substring 's := s is.num s "." = or char.num k = not and char.num #1 + 'char.num := while char.num #1 - 'char.num := t #1 char.num substring FUNCTI...

  43. [51]

    sn-nature.bst

    FUNCTION identify.nature.version "sn-nature.bst" " [2024/07/19 v1.1 bibliography style]" * top ENTRY address archive author booktitle chapter edition editor eprint howpublished institution journal key keywords month note number organization pages publisher school series title ...

    2024

  44. [52]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  45. [53]

    , Nocedal , J

    barticle Liu , D.C. , Nocedal , J. : On the limited memory BFGS method for large scale optimization . Math. Program. 45 ( 1-3 ), 503 -- 528 ( 1989 ) 10.1007/BF01589116 barticle

    1989 doi

  46. [54]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  47. [55]

    sn-vancouver-num.bst

    FUNCTION identify.vancouver.version "sn-vancouver-num.bst" " [2024/07/19 v1.1 Vancouver bibliography style]" * top ENTRY address assignee author booktitle chapter cartographer day edition editor howpublished institution inventor journal key keywords month note number organizat...

    2024

  48. [56]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.