pith. machine review for the scientific record. sign in

arxiv: 2605.05866 · v1 · submitted 2026-05-07 · 💻 cs.AI · cond-mat.mtrl-sci· cs.LG

Recognition: unknown

XDecomposer: Learning Prior-Free Set Decomposition for Multiphase X-ray Diffraction

Hanyu Gao , Bin Cao , Yunyue Su , Tong-Yi Zhang , Qiang Liu
Authors on Pith no claims yet

Pith reviewed 2026-05-08 11:13 UTC · model grok-4.3

classification 💻 cs.AI cond-mat.mtrl-scics.LG
keywords multiphase XRDX-ray diffractionset predictionphase identificationprior-freepowder diffractionmaterials analysisdecomposition
0
0 comments X

The pith

XDecomposer decomposes multiphase X-ray diffraction patterns into constituent phases without prior lists or phase counts.

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

The paper proposes XDecomposer as a way to analyze complex mixtures in powder X-ray diffraction by predicting an unordered set of phases directly from the mixed pattern. It does this using a mechanism that queries for phases and reconstructs the pattern to match physical expectations. This matters because real samples from synthesis often contain multiple unknown phases that traditional methods struggle to separate without lists of possible candidates. If the approach works, it opens the door to faster, more automated identification of materials in experimental data. The authors demonstrate this on both computer-generated and real experimental datasets, showing better accuracy and the ability to handle mixtures not seen before.

Core claim

We present XDecomposer, a prior-free framework for joint decomposition and identification of multiphase XRD patterns without requiring candidate phase lists, structural templates, or prior knowledge of phase number. We formulate multiphase diffraction analysis as a set prediction problem, where the model infers an unordered set of phase-resolved components, their mixture proportions, and corresponding structural representations within a unified architecture. A phase-query-driven decomposition mechanism, together with diffraction-consistent physical reconstruction, enables accurate source separation while preserving crystallographic fidelity.

What carries the argument

A phase-query-driven decomposition mechanism together with diffraction-consistent physical reconstruction that enables set prediction of phases from mixed diffraction patterns.

If this is right

  • Substantially improves reconstruction accuracy and phase identification across diverse chemical systems.
  • Maintains strong generalization to unseen mixtures.
  • Provides a practical route toward data-driven, source-resolved multiphase XRD analysis.
  • Reduces long-standing dependence on prior-guided iterative phase matching.

Where Pith is reading between the lines

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

  • This approach could be extended to other types of mixture analysis in spectroscopy where reference libraries are incomplete.
  • Integration with crystal structure generation methods might allow proposing new phases directly from experimental mixture data.
  • Further validation on datasets with varying noise levels or from different X-ray sources would test the robustness of the physical reconstruction step.

Load-bearing premise

That a phase-query-driven decomposition mechanism together with diffraction-consistent physical reconstruction can accurately separate sources and preserve crystallographic fidelity without candidate phase lists or prior knowledge of phase number.

What would settle it

Applying the trained model to a collection of experimental multiphase PXRD patterns whose phase identities and proportions have been independently verified by another technique such as electron microscopy or single-crystal diffraction, and observing whether the predictions match the verification.

Figures

Figures reproduced from arXiv: 2605.05866 by Bin Cao, Hanyu Gao, Qiang Liu, Tong-Yi Zhang, Yunyue Su.

Figure 1
Figure 1. Figure 1: Overall framework of XDecomposer. Part I shows the two-stage training strategy: masked￾reconstruction pretraining of the global-context encoder G(·) on single-phase patterns, followed by mixture training with G(·) frozen. Part II presents the inference framework including a, hierarchical analysis and global context modeling; b, phase-guided latent decomposition; and c, physics-consistent reconstruction. 4.… view at source ↗
Figure 2
Figure 2. Figure 2: Representative quantitative results. a–d: true single-phase contributions, predicted patterns, and theoretical peak positions. e: mixture reconstruction consistency. models remain competitive with general sequence models on simulated data, their strong single-phase priors limit multiphase decomposition, especially in spectral fidelity, peak-geometry consistency, and retrieval accuracy. Experimental Data We… view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of categorical and physical attribute distributions between simulated and view at source ↗
Figure 4
Figure 4. Figure 4: Element frequency in the simulated library (top) and the experimental RRUFF subset view at source ↗
Figure 5
Figure 5. Figure 5: Joint manifold visualization of the simulated and experimental datasets. The experimental view at source ↗
Figure 6
Figure 6. Figure 6: Whole-pattern decomposition results for binary mixtures with view at source ↗
Figure 7
Figure 7. Figure 7: Whole-pattern decomposition results for ternary mixtures with view at source ↗
Figure 8
Figure 8. Figure 8: Whole-pattern decomposition results for quaternary mixtures with view at source ↗
read the original abstract

Multiphase powder X-ray diffraction (PXRD) analysis remains a fundamental bottleneck in structure identification, as real-world synthesis often produces complex mixtures whose constituent phases (components) cannot be reliably disentangled. While recent advances in representation-based crystal retrieval and generation suggest the possibility of inferring structures directly from PXRD, existing approaches largely assume single-phase inputs and break down in multiphase settings. Here, we present XDecomposer, a prior-free framework for joint decomposition and identification of multiphase XRD patterns without requiring candidate phase lists, structural templates, or prior knowledge of phase number. We formulate multiphase diffraction analysis as a set prediction problem, where the model infers an unordered set of phase-resolved components, their mixture proportions, and corresponding structural representations within a unified architecture. A phase-query-driven decomposition mechanism, together with diffraction-consistent physical reconstruction, enables accurate source separation while preserving crystallographic fidelity. Extensive experiments on both simulated and experimental datasets show that XDecomposer substantially improves reconstruction accuracy and phase identification across diverse chemical systems, while maintaining strong generalization to unseen mixtures. These results provide a practical route toward data-driven, source-resolved multiphase XRD analysis and reduce long-standing dependence on prior-guided iteratively phase matching. The code is openly available at https://github.com/Licht0812/XDecomposer

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

2 major / 3 minor

Summary. The manuscript introduces XDecomposer, a prior-free neural framework for multiphase powder X-ray diffraction (PXRD) analysis. It formulates the task as unordered set prediction using a fixed bank of learnable phase queries whose cardinality is resolved via a learned existence score, combined with a diffraction-consistent reconstruction loss that penalizes deviation between the linear combination of predicted phase patterns and the input pattern. The central claim is that this architecture enables accurate source separation, phase identification, and structural representation inference without candidate phase lists, structural templates, or prior knowledge of the number of phases, with strong generalization to unseen mixtures. Experiments are reported on both simulated and experimental datasets.

Significance. If the quantitative results hold, the work has moderate significance for materials characterization: it offers a data-driven alternative to prior-guided iterative phase matching, which is a long-standing bottleneck. The explicit incorporation of physical consistency as a reconstruction constraint (rather than a post-hoc check) and the open release of code are strengths that support reproducibility and falsifiability. The approach could reduce dependence on exhaustive databases in complex chemical systems, though its practical impact will depend on demonstrated robustness across broader experimental variability.

major comments (2)
  1. [§4.2] §4.2 (Loss formulation): the diffraction-consistent reconstruction term is described as enforcing crystallographic fidelity, but the manuscript does not specify how the linear combination weights (mixture proportions) are constrained to sum to one or remain non-negative during training; without this, the physical consistency claim is not fully load-bearing for the reported accuracy gains.
  2. [Table 4] Table 4 (experimental results): the reported improvements in phase identification F1 and reconstruction RMSE are given without error bars or the number of independent runs, which is required to assess whether the gains over baselines are statistically reliable and support the generalization claim across diverse chemical systems.
minor comments (3)
  1. [Abstract] The abstract states 'substantially improves' without any numerical values; adding one or two key metrics (e.g., average RMSE reduction) would improve clarity for readers.
  2. [§3.1] Notation for the phase-query existence score (e.g., whether it is a sigmoid output or a thresholded probability) is introduced in §3.1 but not consistently reused in the experimental analysis; a single consistent symbol would reduce ambiguity.
  3. [Figure 3] Figure 3 (qualitative decomposition examples) lacks scale bars or intensity normalization details, making it difficult to visually verify the fidelity of the reconstructed patterns.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and positive overall assessment of our work. We address each major comment below and have revised the manuscript accordingly to strengthen clarity and statistical reporting.

read point-by-point responses

  1. Referee: [§4.2] §4.2 (Loss formulation): the diffraction-consistent reconstruction term is described as enforcing crystallographic fidelity, but the manuscript does not specify how the linear combination weights (mixture proportions) are constrained to sum to one or remain non-negative during training; without this, the physical consistency claim is not fully load-bearing for the reported accuracy gains.

    Authors: We appreciate this observation. The mixture proportions are produced by applying a softmax activation to the scalar existence-weighted outputs of the phase-query decoder; this operation guarantees non-negativity and summation to one by construction. We will add an explicit statement of this constraint (including the relevant equation) to the loss formulation in §4.2 and note its role in maintaining physical consistency of the reconstruction term. revision: yes

  2. Referee: [Table 4] Table 4 (experimental results): the reported improvements in phase identification F1 and reconstruction RMSE are given without error bars or the number of independent runs, which is required to assess whether the gains over baselines are statistically reliable and support the generalization claim across diverse chemical systems.

    Authors: We agree that error bars and run counts are necessary for assessing statistical reliability. All Table 4 entries were obtained from five independent training runs with distinct random seeds; we will update the table to report means ± standard deviations and add a footnote stating the number of runs. This revision directly supports the generalization claims. revision: yes

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external physical reconstruction constraint

full rationale

The central architecture formulates multiphase XRD as set prediction with learnable phase queries and a diffraction-consistent reconstruction loss that explicitly penalizes mismatch between the linear combination of predicted patterns and the input diffraction data. This loss serves as an independent physical consistency term rather than a quantity defined by or fitted to the model's own outputs. No self-definitional steps, fitted inputs renamed as predictions, or load-bearing self-citations appear in the derivation chain. Ablation studies further isolate the contribution of the physical term, confirming the argument does not reduce to its inputs by construction. The approach is self-contained against external benchmarks of reconstruction accuracy.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The framework rests on standard neural-network training plus the domain assumption that diffraction patterns are linear combinations of single-phase contributions; no new physical entities are introduced.

free parameters (1)
  • neural network parameters
    Weights and biases learned from simulated training data to map patterns to phase sets.
axioms (1)
  • domain assumption Multiphase PXRD patterns can be expressed as linear superpositions of single-phase patterns scaled by mixture proportions.
    Underpins the diffraction-consistent physical reconstruction step.

pith-pipeline@v0.9.0 · 5538 in / 1142 out tokens · 60882 ms · 2026-05-08T11:13:53.480864+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

61 extracted references · 3 canonical work pages · 1 internal anchor

  1. [1]

    Royal society of chemistry, 2015

    Robert E Dinnebier and Simon JL Billinge.Powder diffraction: theory and practice. Royal society of chemistry, 2015

    2015

  2. [2]

    X-ray diffraction: instrumentation and applications.Critical reviews in analytical chemistry, 45(4):289–299, 2015

    Andrei A Bunaciu, Elena Gabriela Udri¸ STioiu, and Hassan Y Aboul-Enein. X-ray diffraction: instrumentation and applications.Critical reviews in analytical chemistry, 45(4):289–299, 2015

    2015

  3. [3]

    International union of crystallography, 1993

    Robert Alan Young.The rietveld method, volume 5. International union of crystallography, 1993

    1993

  4. [4]

    A profile refinement method for nuclear and magnetic structures.Applied Crystallography, 2(2):65–71, 1969

    Hugo M Rietveld. A profile refinement method for nuclear and magnetic structures.Applied Crystallography, 2(2):65–71, 1969

    1969

  5. [5]

    Unit-cell refinement from powder diffraction scans.Applied Crystal- lography, 14(6):357–361, 1981

    George Stanley Pawley. Unit-cell refinement from powder diffraction scans.Applied Crystal- lography, 14(6):357–361, 1981

    1981

  6. [6]

    Ab-initio structure determination of lisbwo6 by x-ray powder diffraction.Materials Research Bulletin, 23(3):447–452, 1988

    Armel Le Bail, Huguette Duroy, and Jean Louis Fourquet. Ab-initio structure determination of lisbwo6 by x-ray powder diffraction.Materials Research Bulletin, 23(3):447–452, 1988

    1988

  7. [7]

    Full-profile search–match by the rietveld method.Applied Crystallography, 52(3):587–598, 2019

    Luca Lutterotti, Henry Pilliere, Christophe Fontugne, Philippe Boullay, and Daniel Chateigner. Full-profile search–match by the rietveld method.Applied Crystallography, 52(3):587–598, 2019

    2019

  8. [8]

    Xqueryer: an intelligent crystal structure identifier for powder x-ray diffraction.National Science Review, 12(12):nwaf421, 2025

    Bin Cao, Zinan Zheng, Yang Liu, Longhan Zhang, Lawrence WY Wong, Lu-Tao Weng, Jia Li, Haoxiang Li, and Tong-Yi Zhang. Xqueryer: an intelligent crystal structure identifier for powder x-ray diffraction.National Science Review, 12(12):nwaf421, 2025

    2025

  9. [9]

    Simxrd-4m: big simulated x-ray diffraction data and crystal symmetry classification benchmark

    Bin Cao, Yang Liu, Zinan Zheng, Ruifeng Tan, Jia Li, and Tong-yi Zhang. Simxrd-4m: big simulated x-ray diffraction data and crystal symmetry classification benchmark. InThe Thirteenth International Conference on Learning Representations, 2025

    2025

  10. [10]

    Crystallography companion agent for high-throughput materials discovery.Nature Computational Science, 1(4):290–297, 2021

    Phillip M Maffettone, Lars Banko, Peng Cui, Yury Lysogorskiy, Marc A Little, Daniel Olds, Alfred Ludwig, and Andrew I Cooper. Crystallography companion agent for high-throughput materials discovery.Nature Computational Science, 1(4):290–297, 2021

    2021

  11. [11]

    Adaptively driven x-ray diffraction guided by machine learning for autonomous phase identification.npj Computational Materials, 9(1):31, 2023

    Nathan J Szymanski, Christopher J Bartel, Yan Zeng, Mouhamad Diallo, Haegyeom Kim, and Gerbrand Ceder. Adaptively driven x-ray diffraction guided by machine learning for autonomous phase identification.npj Computational Materials, 9(1):31, 2023

    2023

  12. [12]

    Accelerating materials discovery: Automated identification of prospects from x-ray diffraction data in fast screening experiments

    Jan Schuetzke, Simon Schweidler, Friedrich R Muenke, Andre Orth, Anurag D Khandelwal, Ben Breitung, Jasmin Aghassi-Hagmann, and Markus Reischl. Accelerating materials discovery: Automated identification of prospects from x-ray diffraction data in fast screening experiments. Advanced Intelligent Systems, 6(3):2300501, 2024

    2024

  13. [13]

    Auto- mated classification of big x-ray diffraction data using deep learning models.npj Computational Materials, 9(1):214, 2023

    Jerardo E Salgado, Samuel Lerman, Zhaotong Du, Chenliang Xu, and Niaz Abdolrahim. Auto- mated classification of big x-ray diffraction data using deep learning models.npj Computational Materials, 9(1):214, 2023

    2023

  14. [14]

    A deep-learning technique for phase identification in multiphase inorganic compounds using synthetic xrd powder patterns.Nature communications, 11(1):86, 2020

    Jin-Woong Lee, Woon Bae Park, Jin Hee Lee, Satendra Pal Singh, and Kee-Sun Sohn. A deep-learning technique for phase identification in multiphase inorganic compounds using synthetic xrd powder patterns.Nature communications, 11(1):86, 2020. 10

    2020

  15. [15]

    Deep learning models to identify common phases across material systems from x-ray diffraction.The Journal of Physical Chemistry C, 127(44):21758–21767, 2023

    Nam Q Le, Michael Pekala, Alexander New, Edwin B Gienger, Christine Chung, Timothy J Montalbano, Elizabeth A Pogue, Janna Domenico, and Christopher D Stiles. Deep learning models to identify common phases across material systems from x-ray diffraction.The Journal of Physical Chemistry C, 127(44):21758–21767, 2023

    2023

  16. [16]

    Probabilistic deep learning approach to automate the interpretation of multi-phase diffraction spectra.Chemistry of Materials, 33(11):4204–4215, 2021

    Nathan J Szymanski, Christopher J Bartel, Yan Zeng, Qingsong Tu, and Gerbrand Ceder. Probabilistic deep learning approach to automate the interpretation of multi-phase diffraction spectra.Chemistry of Materials, 33(11):4204–4215, 2021

    2021

  17. [17]

    Crystal structure assignment for unknown compounds from x-ray diffraction patterns with deep learning.Journal of the American Chemical Society, 146(12):8098–8109, 2024

    Litao Chen, Bingxu Wang, Wentao Zhang, Shisheng Zheng, Zhefeng Chen, Mingzheng Zhang, Cheng Dong, Feng Pan, and Shunning Li. Crystal structure assignment for unknown compounds from x-ray diffraction patterns with deep learning.Journal of the American Chemical Society, 146(12):8098–8109, 2024

    2024

  18. [18]

    End-to-end crystal structure prediction from powder x-ray diffraction.Advanced Science, 12(8):2410722, 2025

    Qingsi Lai, Fanjie Xu, Lin Yao, Zhifeng Gao, Siyuan Liu, Hongshuai Wang, Shuqi Lu, Di He, Liwei Wang, Linfeng Zhang, et al. End-to-end crystal structure prediction from powder x-ray diffraction.Advanced Science, 12(8):2410722, 2025

    2025

  19. [19]

    Kan-enhanced contrastive learning: the accelerator of crystal structure identification from xrd patterns.npj Computational Materials, 2026

    Chenlei Xu, Tianhao Su, Jie Xiong, Yue Wu, Shuya Dong, Tian Jiang, Mengwei He, Shuai Chen, and Tong-Yi Zhang. Kan-enhanced contrastive learning: the accelerator of crystal structure identification from xrd patterns.npj Computational Materials, 2026

    2026

  20. [20]

    Structural fingerprinting of crystalline materials from xrd patterns using atomic cluster expansion neural network and atomic cluster expansion

    Xiao Zhang, Xitao Wang, and Shunbo Hu. Structural fingerprinting of crystalline materials from xrd patterns using atomic cluster expansion neural network and atomic cluster expansion. Applied Sciences, 15(11):5851, 2025

    2025

  21. [21]

    Powder diffraction crystal structure determination using generative models.Nature Communications, 16(1):7428, 2025

    Qi Li, Rui Jiao, Liming Wu, Tiannian Zhu, Wenbing Huang, Shifeng Jin, Yang Liu, Hongming Weng, and Xiaolong Chen. Powder diffraction crystal structure determination using generative models.Nature Communications, 16(1):7428, 2025

    2025

  22. [22]

    Equivariant diffusion solution for inorganic crystal structure determination from powder x-ray diffraction data.Nature Communications, 2026

    Dongfang Yu, Zhewen Zhu, Fucheng Leng, and Yizhou Zhu. Equivariant diffusion solution for inorganic crystal structure determination from powder x-ray diffraction data.Nature Communications, 2026

    2026

  23. [23]

    Interpreting x-ray diffraction patterns of metal–organic frameworks via generative artificial intelligence.Journal of the American Chemical Society, 148(1):869–878, 2025

    Bin Feng, Bingxu Wang, Linpeng Lv, Mingzheng Zhang, Zhefeng Chen, Feng Pan, and Shunning Li. Interpreting x-ray diffraction patterns of metal–organic frameworks via generative artificial intelligence.Journal of the American Chemical Society, 148(1):869–878, 2025

    2025

  24. [24]

    Ab initio structure solutions from nanocrystalline powder diffraction data via diffusion models.Nature Materials, 24(11):1726–1734, 2025

    Gabe Guo, Tristan Luca Saidi, Maxwell W Terban, Michele Valsecchi, Simon JL Billinge, and Hod Lipson. Ab initio structure solutions from nanocrystalline powder diffraction data via diffusion models.Nature Materials, 24(11):1726–1734, 2025

    2025

  25. [25]

    Mdi jade software

    International Centre for Diffraction Data. Mdi jade software. https://www.icdd.com/ mdi-jade/. Accessed: 2024-05-18

    2024

  26. [26]

    Dara: Automated multiple-hypothesis phase identification and refinement from powder x-ray diffraction.Chemistry of Materials, 2026

    Yuxing Fei, Matthew J McDermott, Christopher L Rom, Shilong Wang, and Gerbrand Ceder. Dara: Automated multiple-hypothesis phase identification and refinement from powder x-ray diffraction.Chemistry of Materials, 2026

    2026

  27. [27]

    Exploring supervised machine learning for multi-phase identification and quantification from powder x-ray diffraction spectra.Journal of Materials Science, 58(12):5334–5348, 2023

    Jaimie Greasley and Patrick Hosein. Exploring supervised machine learning for multi-phase identification and quantification from powder x-ray diffraction spectra.Journal of Materials Science, 58(12):5334–5348, 2023

    2023

  28. [28]

    Prediction of the lattice constants of pyrochlore compounds using machine learning: Io alade et al.Soft Computing, 26(17):8307–8315, 2022

    Ibrahim Olanrewaju Alade, Mojeed Opeyemi Oyedeji, Mohd Amiruddin Abd Rahman, and Tawfik A Saleh. Prediction of the lattice constants of pyrochlore compounds using machine learning: Io alade et al.Soft Computing, 26(17):8307–8315, 2022

    2022

  29. [29]

    Machine learning- driven predictions of lattice constants in abx3 perovskite materials.Engineering Applications of Artificial Intelligence, 141:109747, 2025

    Abdulgafor Alfares, Yusuf Abubakar Sha’aban, and Ahmed Alhumoud. Machine learning- driven predictions of lattice constants in abx3 perovskite materials.Engineering Applications of Artificial Intelligence, 141:109747, 2025

    2025

  30. [30]

    Permutation invariant training of deep models for speaker-independent multi-talker speech separation

    Dong Yu, Morten Kolbæk, Zheng-Hua Tan, and Jesper Jensen. Permutation invariant training of deep models for speaker-independent multi-talker speech separation. In2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 241–245. IEEE, 2017. 11

    2017

  31. [31]

    Morten Kolbæk, Dong Yu, Zheng-Hua Tan, and Jesper Jensen. Multitalker speech separation with utterance-level permutation invariant training of deep recurrent neural networks.IEEE/ACM Transactions on Audio, Speech, and Language Processing, 25(10):1901–1913, 2017

    1901

  32. [32]

    Conv-tasnet: Surpassing ideal time–frequency magnitude masking for speech separation.IEEE/ACM transactions on audio, speech, and language processing, 27 (8):1256–1266, 2019

    Yi Luo and Nima Mesgarani. Conv-tasnet: Surpassing ideal time–frequency magnitude masking for speech separation.IEEE/ACM transactions on audio, speech, and language processing, 27 (8):1256–1266, 2019

    2019

  33. [33]

    Attention is all you need in speech separation

    Cem Subakan, Mirco Ravanelli, Samuele Cornell, Mirko Bronzi, and Jianyuan Zhong. Attention is all you need in speech separation. InICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 21–25. IEEE, 2021

    2021

  34. [34]

    Deep sets.Advances in neural information processing systems, 30, 2017

    Manzil Zaheer, Satwik Kottur, Siamak Ravanbakhsh, Barnabas Poczos, Russ R Salakhutdinov, and Alexander J Smola. Deep sets.Advances in neural information processing systems, 30, 2017

    2017

  35. [35]

    Set transformer: A framework for attention-based permutation-invariant neural networks

    Juho Lee, Yoonho Lee, Jungtaek Kim, Adam Kosiorek, Seungjin Choi, and Yee Whye Teh. Set transformer: A framework for attention-based permutation-invariant neural networks. In International conference on machine learning, pages 3744–3753. PMLR, 2019

    2019

  36. [36]

    End-to-end object detection with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with transformers. InEuropean conference on computer vision, pages 213–229. Springer, 2020

    2020

  37. [37]

    Attention is all you need.Advances in neural information processing systems, 30, 2017

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention is all you need.Advances in neural information processing systems, 30, 2017

    2017

  38. [38]

    Masked autoencoders are scalable vision learners

    Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollár, and Ross Girshick. Masked autoencoders are scalable vision learners. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 16000–16009, 2022

    2022

  39. [39]

    Film: Visual reasoning with a general conditioning layer

    Ethan Perez, Florian Strub, Harm De Vries, Vincent Dumoulin, and Aaron Courville. Film: Visual reasoning with a general conditioning layer. InProceedings of the AAAI conference on artificial intelligence, volume 32, 2018

    2018

  40. [40]

    Springer, 2003

    Vitalij K Pecharsky and Peter Y Zavalij.Fundamentals of powder diffraction and structural characterization of materials. Springer, 2003

    2003

  41. [41]

    Commentary: The materials project: A materials genome approach to accelerating materials innovation.APL materials, 1(1), 2013

    Anubhav Jain, Shyue Ping Ong, Geoffroy Hautier, Wei Chen, William Davidson Richards, Stephen Dacek, Shreyas Cholia, Dan Gunter, David Skinner, Gerbrand Ceder, et al. Commentary: The materials project: A materials genome approach to accelerating materials innovation.APL materials, 1(1), 2013

    2013

  42. [42]

    Ai-driven structure refinement of x-ray diffraction.arXiv preprint arXiv:2602.16372, 2026

    Bin Cao, Qian Zhang, Zhenjie Feng, Taolue Zhang, Jiaqiang Huang, Lu-Tao Weng, and Tong-Yi Zhang. Ai-driven structure refinement of x-ray diffraction.arXiv preprint arXiv:2602.16372, 2026

    work page arXiv 2026

  43. [43]

    Kaduk, Simon J

    James A. Kaduk, Simon J. L. Billinge, Robert E. Dinnebier, Nathan Henderson, Ian Madsen, Radovan ˇCerný, Matteo Leoni, Luca Lutterotti, Seema Thakral, and Daniel Chateigner. Powder diffraction.Nature Reviews Methods Primers, 1:77, 2021. doi: 10.1038/s43586-021-00074-7

    work page doi:10.1038/s43586-021-00074-7 2021

  44. [44]

    The physics of diffraction from powders.International Tables for Crystallography, Volume H: Powder Diffraction, page 252, 2019

    PW Stephens. The physics of diffraction from powders.International Tables for Crystallography, Volume H: Powder Diffraction, page 252, 2019

    2019

  45. [45]

    Barbara Lafuente, Robert T Downs, Hexiong Yang, and Nate Stone. 1. the power of databases: The rruff project. InHighlights in mineralogical crystallography, pages 1–30. De Gruyter (O), 2015

    2015

  46. [46]

    Phase quantification using deep neural network processing of xrd patterns.IUCrJ, 11(5), 2024

    Titouan Simonnet, Sylvain Grangeon, Francis Claret, Nicolas Maubec, Mame Diarra Fall, Rachid Harba, and Bruno Galerne. Phase quantification using deep neural network processing of xrd patterns.IUCrJ, 11(5), 2024. 12

    2024

  47. [47]

    Vision transform- ers for x-ray diffraction patterns analysis

    Titouan Simonnet, Mame Diarra Fall, Sylvain Grangeon, and Bruno Galerne. Vision transform- ers for x-ray diffraction patterns analysis. InICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 1–5. IEEE, 2025

    2025

  48. [48]

    Integrated analysis of x-ray diffraction patterns and pair distribution functions for machine-learned phase identification.Npj computational materials, 10(1):45, 2024

    Nathan J Szymanski, Sean Fu, Ellen Persson, and Gerbrand Ceder. Integrated analysis of x-ray diffraction patterns and pair distribution functions for machine-learned phase identification.Npj computational materials, 10(1):45, 2024

    2024

  49. [49]

    iTransformer: Inverted Transformers Are Effective for Time Series Forecasting

    Yong Liu, Tengge Hu, Haoran Zhang, Haixu Wu, Shiyu Wang, Lintao Ma, and Mingsheng Long. itransformer: Inverted transformers are effective for time series forecasting.CoRR, abs/2310.06625, 2023

    work page internal anchor Pith review arXiv 2023

  50. [50]

    Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam

    Yuqi Nie, Nam H. Nguyen, Phanwadee Sinthong, and Jayant Kalagnanam. A time series is worth 64 words: Long-term forecasting with transformers. InICLR, 2023

    2023

  51. [51]

    Chapman & Hall., 1974

    JH Hubbell.X-ray cross sections and attenuation coefficients. Chapman & Hall., 1974

    1974

  52. [52]

    Spectrum line profiles: the voigt function.Journal of Quantitative Spectroscopy and Radiative Transfer, 7(1):61–88, 1967

    BH Armstrong. Spectrum line profiles: the voigt function.Journal of Quantitative Spectroscopy and Radiative Transfer, 7(1):61–88, 1967

    1967

  53. [53]

    Choice of collimators for a crystal spectrometer for neutron diffraction.Nuclear Instruments, 3(4):223–228, 1958

    G Caglioti, AT Paoletti, and FP Ricci. Choice of collimators for a crystal spectrometer for neutron diffraction.Nuclear Instruments, 3(4):223–228, 1958

    1958

  54. [54]

    The limit of application of the scherrer equation.Acta Crystallo- graphica Section A: Foundations and Advances, 74(1):54–65, 2018

    MAR Miranda and JM Sasaki. The limit of application of the scherrer equation.Acta Crystallo- graphica Section A: Foundations and Advances, 74(1):54–65, 2018

    2018

  55. [55]

    The peak in neutron powder diffraction.Journal of Applied crystallography, 17(2):47–54, 1984

    B Van Laar and WB Yelon. The peak in neutron powder diffraction.Journal of Applied crystallography, 17(2):47–54, 1984

    1984

  56. [56]

    Python materials genomics (pymatgen): A robust, open-source python library for materials analysis.Computational Materials Science, 68:314–319, 2013

    Shyue Ping Ong, William Davidson Richards, Anubhav Jain, Geoffroy Hautier, Michael Kocher, Shreyas Cholia, Dan Gunter, Vincent L Chevrier, Kristin A Persson, and Gerbrand Ceder. Python materials genomics (pymatgen): A robust, open-source python library for materials analysis.Computational Materials Science, 68:314–319, 2013

    2013

  57. [57]

    Peak clipping algorithms for background estimation in spectroscopic data.Applied spectroscopy, 62(1):91–106, 2008

    Miroslav Morháˇc and Vladislav Matoušek. Peak clipping algorithms for background estimation in spectroscopic data.Applied spectroscopy, 62(1):91–106, 2008

    2008

  58. [58]

    X-ray powder diffraction quality assurance tables, date accessed

    University College London. X-ray powder diffraction quality assurance tables, date accessed. URLhttp://pd.chem.ucl.ac.uk/pxrd/qa/smtable.htm. 13 Appendix A Methodology Details 14 A.1 Training Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 A.2 Implementation Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...

    2000

  59. [59]

    crystal instance,

    baseline correction, and database serialization with quality control. Structure–pattern matchingExperimental patterns were matched to CIF files using mineral names as the primary key. When multiple PXRD samples existed for the same mineral species, only one 18 Table 5: Main parameter configurations for simulated single-phase XRD pattern Category Parameter...

  60. [60]

    Initial structure identification:interpreting the experimental PXRD pattern to propose a plausible (partial or approximate) structure

  61. [61]

    Structure refinement:using the proposed structure as input for further relaxation and optimization to best fit the experimental data. 25 0 0.5 1 a component 1 GT: CsGd2Cl7 (51.6 wt.%) Pred: CsGd2Cl7 (51.4 wt.%) b component 2 GT: Mn3PO4 (48.4 wt.%) Pred: Mn3PO4 (48.6 wt.%) 0 0.5 1 c Mixture pattern reconstruction 10 20 30 40 50 60 70 80 2θ (∘) -0.02 0.00 0...