Pith. sign in

REVIEW 4 major objections 5 minor 41 references

AutoMat: Enabling Automated Crystal Structure Reconstruction from Microscopy via Agentic Tool Use

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read AutoMat claims that a text-only language model, armed with denoising, template retrieval, reconstruction, and relaxation tools, can turn a single noisy STEM micrograph into a simulation-ready crystal structure and a formation-energy…

desk verdict AutoMat is a serious engineering effort with a likely inflated headline: template leakage and inconsistent tier counts undermine the accuracy claims until fixed. read the letter →

arxiv 2505.12650 v2 pith:G2SWBI5C submitted 2025-05-19 cs.CV cs.AI

Yaotian Yang , Yiwen Tang , Yizhe Chen , Xiao Chen , Jiangjie Qiu , Hao Xiong , Haoyu Yin , Zhiyao Luo
show 9 more authors
Yifei Zhang Sijia Tao Wentao Li Qinghua Zhang Yuqiang Li Wanli Ouyang Bin Zhao Xiaonan Wang Fei Wei
This is my paper · ORCID
classification cs.CVcs.AI
keywords STEMimageanalysiscrystalstructurereconstructionagentictooluseclosed-loopverificationmachine-learnedinteratomicpotentialstemplateretrieval2Dmaterialsformationenergyprediction
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 claims that a single noisy STEM image of a 2D crystal can be converted automatically into a simulation-ready crystal structure and a formation-energy estimate, without a human annotator. The proposed system, AutoMat, chains four tools—adaptive denoising, physics-guided template retrieval, symmetry-constrained reconstruction into a CIF, and relaxation plus energy prediction with a machine-learned interatomic potential—and lets a text-only language model orchestrate them with rollback-and-retry when a quality check fails. To test this, the authors build STEM2Mat-Bench, 450 simulated image–structure–energy triples drawn from 2,143 curated 2D materials, and report a projected lattice RMSD of about 0.11 Å, a formation-energy MAE near 330 meV/atom, and an 83.2% structure success rate, an order of magnitude better than vision-language model baselines and specialized STEM toolkits. The paper's point is that closed-loop tool use, not a bigger vision model, is what bridges microscopy and atomistic simulation.

What carries the argument

The load-bearing mechanism is the agentic loop around the four tools: a language-model controller makes tool calls, receives structured intermediate results, runs quality checks, and can roll back to a previous stage and retry. The physics-guided template retrieval branch is the crucial state-dependent auxiliary: rather than interpreting the image ab initio, the system matches the denoised image to a library of simulated STEM projections, which supplies a strong prior for lattice type and element assignment; the reconstruction module then refines the candidate under symmetry constraints. The closed-loop verification is what distinguishes this from a one-pass feed-forward reconstruction.

What would settle it

Run AutoMat's template-matching stage on the STEM2Mat test images after deleting every test structure's ground-truth template from the gallery (or with a gallery of unrelated structures), and check whether the 83.2% success rate and 0.11 Å RMSD collapse toward retrieval-by-chance; also compare against a closed-loop method that receives no template gallery at all.

Watch

Extended reading notes

Core claim

The central claim is that inference-time hypothesis search with closed-loop verification lets an otherwise text-only LLM outperform vision-language models on a spatially precise inverse problem. AutoMat treats reconstruction as a sequence of tool calls: a pattern-adaptive denoiser enhances the micrograph; an image template matcher proposes candidate structures from a library of simulated projections, filtered by elemental contrast; a reconstruction module detects atomic peaks by clustering, fits the lattice under symmetry constraints, assigns species from the candidate, and writes a CIF; and a machine-learned interatomic potential relaxes the structure and predicts formation energy. The agent monitors intermediate outputs and rolls back to retry failed stages. On STEM2Mat-Bench the system reaches 0.11 ± 0.03 Å in-plane lattice RMSD, 321.6 meV/atom mean formation-energy error, and 83.2% structure success, against a 48.1 meV/atom oracle floor when the true CIF is fed directly to the potential. Error analysis attributes 39.3% of failures to template retrieval and 60.7% to downstream steps such as projection ambiguity or element contrast confusion.

Load-bearing premise

The reported accuracy assumes the retrieval library can contain the exact test structure, because the paper does not state that test structures were excluded from the 2,143-template gallery built from the same pool used to create the 450 test images.

Editorial extensions

If this is right

  • If the reported accuracy holds, laboratories could pipe raw STEM micrographs into an automated system that outputs a CIF ready for simulation, removing a bottleneck in training and validating machine-learned interatomic potentials.
  • A text-only LLM equipped with domain tools can beat large vision-language models on a visual, spatially precise scientific task, suggesting that tool orchestration matters more than native image reasoning for such problems.
  • Since 39.3% of failures are attributed to template retrieval, improving retrieval robustness—via uncertainty-aware or multi-candidate matching—would yield the largest immediate gains, while the remaining failures come from projection ambiguity and elemental confusion even with a correct template.
  • Because the paper attributes most residual energy error to reconstruction rather than to the machine-learned potential, structural fidelity improvements should translate almost directly into better property predictions.

Reading between the lines

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

  • The benchmark design leaves the template-gallery overlap question open: the 450 test images are drawn from the same 2,143-structure pool used to build the retrieval library, and the paper never states that test structures are excluded. If they are not, the headline numbers blend retrieval with refinement, and real-world generalization to unseen crystals would likely be lower.
  • The same closed-loop, rollback-on-failure pattern could transfer to other ill-posed imaging inversions, such as 3D atomic reconstruction from a tilt series, where a controller could compare candidate 3D models against multiple projections rather than committing to one feed-forward prediction.
  • A direct extension would be to replace or augment template retrieval with a generative structural head that can hypothesize lattices not present in the library; that would turn AutoMat from a lookup-and-refine system into a true ab initio reconstructor and would resolve the main ambiguity in the evaluation.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper introduces AutoMat, an LLM-orchestrated pipeline that converts a single noisy synthetic STEM projection into a simulation-ready CIF file and a formation-energy prediction. The pipeline chains four modules—pattern-adaptive denoising (MOE-DIV AESR), physics-guided image-template retrieval, symmetry-constrained atomic reconstruction (STEM2CIF), and MLIP-based relaxation/property prediction via MatterSim—with rollback-and-retry orchestrated by a DeepSeek-V3 agent. The authors also introduce STEM2Mat-Bench, a synthetic benchmark of 450 (claimed) image–structure–property triples built from 2,143 curated monolayer structures, and report projected lattice RMSD of about 0.11 Å, formation-energy MAE of about 330 meV/atom, and an overall structure success rate of 83.2%, outperforming GPT-4.1mini, Qwen-VL, LLaMA4V, ChemVLM, and AtomAI by an order of magnitude on the reported tables.

Significance. If the central claims hold, AutoMat would be a useful integration of image denoising, template matching, symmetry-constrained reconstruction, and MLIP validation, and STEM2Mat-Bench would provide a reproducible synthetic evaluation suite for a task that currently lacks standardized benchmarks. The paper has concrete strengths: the authors release code and data, the benchmark construction is described in enough detail to be replicated, the metrics are mostly explicit, and the reported order-of-magnitude improvement over general-purpose VLMs is striking. However, the significance is conditional on two load-bearing points: the test structures must be excluded from the template-retrieval gallery, and the contribution of the LLM agent itself must be separated from the contribution of the specialized modules. The current manuscript does not supply either piece of evidence, and the synthetic-only evaluation further limits the strength of the generalization claims.

major comments (4)
  1. [§3.3, §4.2, §5.2] The manuscript never states that the 450 test ground-truth structures are excluded from the Image Template Matching gallery. Because §3.3 selects the 450 test images from the same 2,143-structure pool used to build the gallery, and §4.2 describes retrieval from 'a pre-stored or externally mounted database of structural templates,' the reported S.S. (83.2%) and RMSD (0.11 Å) could in principle be achieved by retrieving the correct template from the database and locally refining it, rather than by reconstructing the lattice from the image. This is not merely hypothetical: §5.2 attributes 39.3% of failures to template retrieval, confirming that retrieval is load-bearing for the headline result. Please state explicitly whether test structures are excluded from the gallery and, ideally, report a holdout ablation in which the gallery is restricted to the training/validation split.
  2. [§3.3, Tables 1–2] The benchmark size and tier arithmetic are internally inconsistent: the text says 450 test samples were retained, but the tier counts in §3.3 (35 + 456 + 79 = 570) sum to 570, contradicting the abstract, introduction, and §3.3's own '450' figure. Moreover, the aggregate energy MAE of 321.57 in Table 1 matches neither the unweighted mean of the tier means (332.43) nor the weighted mean using the stated tier counts (323.49). Please correct the counts, the aggregate, or the definition of 'Avg.'
  3. [§5.1, Tables 1–2] The central claim is that agentic orchestration by an LLM enables the reported performance, but no ablation compares the DeepSeek-V3 controller against a deterministic, fixed-order execution of the same four modules with the same retry logic. Without such an ablation, the reported improvement over the VLMs cannot be attributed to the agent's closed-loop tool use rather than to the specialized vision and physics modules alone. Please add at least one non-agent ablation (e.g., fixed pipeline, no rollback) and report its tier-wise results.
  4. [§3.4, Eq. (2)] The projected lattice RMSD in Eq. (2) compares only the in-plane lattice-vector lengths a and b and ignores the in-plane angle γ. A prediction with a grossly wrong unit-cell angle can therefore receive a small RMSD, which makes the headline 0.11 Å value incomplete as a structural accuracy measure. Please extend Eq. (2) to include the angle (or the full metric tensor), or justify the current definition and report the angle error separately.
minor comments (5)
  1. [§3.1, §4.2] There are unresolved placeholder references 'Fig. ??' and 'Appendix ?? and ??'; please fill in or remove them before publication.
  2. [Tables 1–2, Figure 4] The model name is written inconsistently as 'LLama4V', 'LLaMA4', and 'Llama-4-Maverick'; please use one canonical name throughout.
  3. [§5.1] The discussion states a mean formation-energy MAE of '332±12 meV/atom', while Table 1 reports an average of 321.57 meV/atom; please reconcile the number and define the reported uncertainty.
  4. [§3.4] The definition of Composition Correctness is per-sample binary, but it is reported as a percentage; please state explicitly that the reported values are means over the test set.
  5. [§3.2, §6] All evaluations use abTEM-simulated images rather than experimental STEM micrographs; please add an explicit limitation statement noting that the benchmark currently measures performance on synthetic data only.

Circularity Check

1 steps flagged · score 7.0 of 10

Benchmark's 450 test structures are drawn from the same 2,143-structure pool used to build the simulated-projection template library, with no stated holdout, so AutoMat's reported reconstruction can reduce to template retrieval plus refinement.

  1. self definitional [Section 3.1/3.3 and Section 4.2]
    "we curated 2,143 high-confidence monolayer crystals ... and simulated their corresponding iDPC-STEM images with abTEM[32]. From this pool we selected 450 representative image–structure pairs, which constitute our STEM2Mat benchmark ... Enhanced images are matched to a large-scale library of simulated STEM projections."

    The 450 test images are selected from the same 2,143-structure pool whose simulated projections are the only described source of the template library. The paper never states that the 450 test structures are excluded from the retrieval database, so for each test image the ground-truth CIF can be a permissible template candidate. STEM2CIF then consumes the selected template, making the output CIF a refined database entry rather than a structure recovered from the image alone. The reported 83.2% structure success and 0.11 Å RMSD therefore measure retrieval-plus-refinement, not ab initio single-image reconstruction. The paper's own error analysis confirms retrieval is load-bearing, attributing 39.3% of failures to template retrieval.

full rationale

The circularity is in the evaluation protocol rather than in the mathematical modules. Energy prediction is checked against an external MLIP (MatterSim) with a ground-truth-CIF upper bound, and the VLM/AtomAI comparisons do not use AutoMat's template gallery, so those parts are independent. The central structural claim, however, is evaluated with a test set carved from the same pool that supplies the retrieval database, and the text never documents that the 450 targets are held out of that database. Because the structural output is explicitly built from the selected template, the headline success rate can be achieved by lookup plus refinement. This makes the central reconstruction claim partially circular: the 'prediction' is not guaranteed to be an inversion of the image, since the answer may already be present in the retrieval input.

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

The central performance claim rests on the fidelity of simulated STEM images, the accuracy of the externally pretrained MatterSim potential, and an unstated assumption about the template gallery not leaking test answers. No new physical entities or free parameters are introduced; the ML models used carry pretrained weights but are not fitted to the test set.

assumptions (4)
  • domain assumption abTEM-simulated iDPC-STEM images with injected Poisson noise and aberrations are representative of real experimental STEM images.
    All 450 test images are synthetic; the paper presents no experimental STEM images, so the transferability of the results rests on this untested equivalence. Invoked in Section 3.2.
  • domain assumption MatterSim's MLIP predictions are accurate enough to serve as the energy oracle and the relaxation engine.
    The paper uses MatterSim for both the upper-bound oracle and AutoMat's energy pipeline, inheriting its pretrained parameters and error characteristics without independent validation here. Invoked in Section 4.2 and Section 5.1.
  • ad hoc to paper The template retrieval gallery does not contain the ground-truth structure for test samples, or if it does, retrieval success does not trivially solve the task.
    The paper never states whether the 450 test structures are excluded from the template library built from the same 2,143-structure pool; the benchmark's validity depends on this unstated condition. Relevant to Section 4.2 and Section 3.3.
  • domain assumption Pixel similarity and elemental contrast suffice to select the correct template among the library candidates.
    The template matching step relies on this similarity heuristic; no evidence is given for its reliability on low-dose ternary images beyond the final end-to-end results. Invoked in Section 4.2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of AutoMat: Enabling Automated Crystal Structure Reconstruction from Microscopy via Agentic Tool Use." pith.science (2026). https://pith.science/paper/G2SWBI5C

@misc{pith2026250512650,
  author       = {Pith},
  title        = {Pith review of: AutoMat: Enabling Automated Crystal Structure Reconstruction from Microscopy via Agentic Tool Use},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G2SWBI5C}},
  note         = {Machine review of arXiv:2505.12650}
}
read the original abstract

Reconstructing atomistic crystal structures from a single noisy STEM projection is an ill-posed inverse problem: multiple lattices can explain similar contrast, and purely feed-forward models cannot verify physical validity. We present AutoMat, a failure-aware agentic controller that performs inference-time hypothesis search with closed-loop verification to convert Scanning Transmission Electron Microscopy (STEM) images into simulation-ready crystal structures and downstream properties. AutoMat composes perception and physics modules---pattern-adaptive denoising, physics-guided template retrieval as a state-dependent auxiliary branch, symmetry-constrained atomic reconstruction, and MLIP-based relaxation/validation---and triggers rollback-and-retry when verification fails. For systematic evaluation, we introduce STEM2Mat-Bench, a benchmark dataset containing 450+ annotated samples. Performance is assessed using lattice root-mean-square deviation (RMSD), formation energy mean absolute error (MAE), and structure matching accuracy. Results demonstrate that AutoMat outperforms existing approaches including SOTA models, specialized domain tools, and closed-source multimodal large models. This work establishes a direct pathway from microscopic characterization to atomic-scale modeling, addressing a fundamental challenge in materials science.

Figures

Figures reproduced from arXiv: 2505.12650 by the authors.

Figure 1
Figure 1. Overview of AutoMat. The left part shows an LLM agent how to manage four modules—pattern-adaptive denoising, template selection, atomic reconstruction (STEM2CIF), and ML-based property prediction. The right panel presents a line chart and radar plot comparing different models in terms of energy and structural errors. structures. Meanwhile, scanning transmission electron microscopy (STEM) can image atoms individ￾uall… view at source ↗
Figure 2
Figure 2. Overview of the STEM2Mat Benchmark design, illustrating the tiered classification of STEM samples by material complexity and imaging dose, which systematically stratifies reconstruc￾tion difficulty from simple unary to complex ternary compounds. we retained only structures containing no more than three distinct elements. For those with multiple elements, we required a minimum atomic-number span of ten, i.e., max(Zi)… view at source ↗
Figure 3
Figure 3. AutoMat’s LLM agent orchestrates four stages—denoising, template matching, structure reconstruction, and property prediction—from STEM image to relaxed crystal with properties. for energy and property predictions. Integrated with ASE, it enables structure optimization and rapid assessment of properties such as total energy, formation energy, and elastic moduli, providing a fast and accurate alternative to convention… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Case studies on Tier 2 and Tier 3 samples comparing AutoMat, GPT-4.1mini and Llama-4- [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 26 canonical work pages

  1. [1]

    Mattersim: A deep learning atomistic model across elements, temperatures and pressures, 2024

    Han Yang, Chenxi Hu, Yichi Zhou, Xixian Liu, Yu Shi, Jielan Li, Guanzhi Li, Zekun Chen, Shuizhou Chen, Claudio Zeni, Matthew Horton, Robert Pinsler, Andrew Fowler, Daniel Zügner, Tian Xie, Jake Smith, Lixin Sun, Qian Wang, Lingyu Kong, Chang Liu, Hongxia Hao, and Ziheng Lu. Mattersim: A deep learning atomistic model across elements, temperatures and press...

    arXiv 2024

  2. [2]

    Elena, Dávid P

    Ilyes Batatia, Philipp Benner, Yuan Chiang, Alin M. Elena, Dávid P. Kovács, Janosh Riebesell, Xavier R. Advincula, Mark Asta, Matthew Avaylon, William J. Baldwin, Fabian Berger, Noam Bernstein, Arghya Bhowmik, Samuel M. Blau, Vlad C ˘arare, James P. Darby, Sandip De, Flaviano Della Pia, V olker L. Deringer, Rokas Elijošius, Zakariya El-Machachi, Fabio Fal...

  3. [3]

    Gemnet: Univer- sal directional graph neural networks for molecules

    Johannes Gasteiger, Florian Becker, and Stephan Günnemann. Gemnet: Univer- sal directional graph neural networks for molecules. In M. Ranzato, A. Beygelz- imer, Y . Dauphin, P.S. Liang, and J. Wortman Vaughan, editors, Advances in Neu- ral Information Processing Systems , volume 34, pages 6790–6802. Curran Associates, Inc., 2021. URL https://proceedings.n...

    work page 2021

  4. [4]

    Equiformerv2: Improved equivariant transformer for scaling to higher-degree representations, 2024

    Yi-Lun Liao, Brandon Wood, Abhishek Das, and Tess Smidt. Equiformerv2: Improved equivariant transformer for scaling to higher-degree representations, 2024. URL https:// arxiv.org/abs/2306.12059

    arXiv 2024

  5. [5]

    Highly accurate quantum chemical property prediction with uni-mol+, 2023

    Shuqi Lu, Zhifeng Gao, Di He, Linfeng Zhang, and Guolin Ke. Highly accurate quantum chemical property prediction with uni-mol+, 2023. URL https://arxiv.org/abs/2303. 16982

    work page 2023

  6. [6]

    Atomic-resolution transmission electron microscopy of electron beam–sensitive crystalline materials

    Daliang Zhang, Yihan Zhu, Lingmei Liu, Xiangrong Ying, Chia-En Hsiung, Rachid Sougrat, Kun Li, and Yu Han. Atomic-resolution transmission electron microscopy of electron beam–sensitive crystalline materials. Science, 359(6376):675–679, 2018. ISSN 0036-8075. 10

    work page 2018

  7. [7]

    Three-dimensional inhomogeneity of zeolite structure and composition revealed by electron ptychography

    Hui Zhang, Guanxing Li, Jiaxing Zhang, Daliang Zhang, Zhen Chen, Xiaona Liu, Peng Guo, Yihan Zhu, Cailing Chen, and Lingmei Liu. Three-dimensional inhomogeneity of zeolite structure and composition revealed by electron ptychography. Science, 380(6645):633–638,

    work page

  8. [8]

    Electron microscopy studies of local structural modulations in zeolite crystals

    Qing Zhang, Alvaro Mayoral, Junyan Li, Juanfang Ruan, Viveka Alfredsson, Yanhang Ma, Jihong Yu, and Osamu Terasaki. Electron microscopy studies of local structural modulations in zeolite crystals. Angewandte Chemie, 132(44):19571–19581, 2020. ISSN 0044-8249

    work page 2020

Show all 41 references
  1. [9]

    Direct imaging of single metal atoms and clusters in the pores of dealuminated hy zeolite

    V olkan Ortalan, Alper Uzun, Bruce C Gates, and Nigel D Browning. Direct imaging of single metal atoms and clusters in the pores of dealuminated hy zeolite. Nature nanotechnology, 5(7): 506–510, 2010. ISSN 1748-3387

    2010

  2. [10]

    Machine learning in scanning transmission electron microscopy

    Sergei V Kalinin, Colin Ophus, Paul M V oyles, Rolf Erni, Demie Kepaptsoglou, Vincenzo Grillo, Andrew R Lupini, Mark P Oxley, Eric Schwenker, and Maria KY Chan. Machine learning in scanning transmission electron microscopy. Nature Reviews Methods Primers, 2(1): 11, 2022. ISSN ...

    2022

  3. [11]

    Machine learning for automated experimentation in scanning transmission electron microscopy

    Sergei V Kalinin, Debangshu Mukherjee, Kevin Roccapriore, Benjamin J Blaiszik, Ayana Ghosh, Maxim A Ziatdinov, Anees Al-Najjar, Christina Doty, Sarah Akers, Nageswara S Rao, et al. Machine learning for automated experimentation in scanning transmission electron microscopy. npj...

    2023

  4. [12]

    Deep learning-enabled stem imaging for precise single-molecule identification in zeo- lite structures

    Yaotian Yang, Hao Xiong, Zirong Wu, Zhiyao Luo, Xiao Chen, Xiaonan Wang, and Fei Wei. Deep learning-enabled stem imaging for precise single-molecule identification in zeo- lite structures. Advanced Science, 12(6):2408629, 2025. doi: https://doi.org/10.1002/advs. 202408629. URL...

    2025 doi

  5. [13]

    Ruoqian Lin, Rui Zhang, Chunyang Wang, Xiao-Qing Yang, and Huolin L Xin. Temimagenet and atomsegnet deep learning training library and models for high-precision atom segmentation, localization, denoising, and super-resolution processing of atom-resolution scanning tem images. ...

    2012 arXiv

  6. [14]

    Temimagenet training library and atomsegnet deep-learning models for high-precision atom segmentation, localization, denoising, and deblurring of atomic-resolution images

    Ruoqian Lin, Rui Zhang, Chunyang Wang, Xiao-Qing Yang, and Huolin L Xin. Temimagenet training library and atomsegnet deep-learning models for high-precision atom segmentation, localization, denoising, and deblurring of atomic-resolution images. Scientific reports, 11(1): 5386,...

    2021

  7. [15]

    Noise2atom: unsupervised denoising for scanning transmission electron microscopy images

    Feng Wang, Trond R Henninen, Debora Keller, and Rolf Erni. Noise2atom: unsupervised denoising for scanning transmission electron microscopy images. Applied Microscopy, 50(1): 23, 2020. ISSN 2287-4445

    2020

  8. [16]

    Ziatdinov, O

    M. Ziatdinov, O. Dyck, A. Maksov, X. Li, X. Sang, K. Xiao, R. R. Unocic, R. Vasudevan, S. Jesse, and S. V . Kalinin. Deep learning of atomically resolved scanning transmission electron microscopy images: Chemical identification and tracking local transformations. ACS Nano, 11 ...

    2017

  9. [17]

    Predicting column heights and elemental compo- sition in experimental transmission electron microscopy images of high-entropy oxides using deep learning

    Ishraque Zaman Borshon, Marco Ragone, Abhijit H Phakatkar, Lance Long, Reza Shahbazian- Yassar, Farzad Mashayek, and Vitaliy Yurkiv. Predicting column heights and elemental compo- sition in experimental transmission electron microscopy images of high-entropy oxides using deep ...

    2024

  10. [18]

    Localization and segmentation of atomic columns in supported nanoparticles for fast scanning transmission electron microscopy

    Henrik Eliasson and Rolf Erni. Localization and segmentation of atomic columns in supported nanoparticles for fast scanning transmission electron microscopy. npj Computational Materials, 10(1):168, 2024. 11

    2024

  11. [19]

    Experimental reconstructions of 3d atomic structures from electron microscopy images using a bayesian genetic algorithm

    Annick De Backer, Sandra Van Aert, Christel Faes, Ece Arslan Irmak, Peter D Nellist, and Lewys Jones. Experimental reconstructions of 3d atomic structures from electron microscopy images using a bayesian genetic algorithm. npj Computational Materials, 8(1):216, 2022

    2022

  12. [20]

    Obtaining 3d atomic recon- structions from electron microscopy images using a bayesian genetic algorithm: Possibilities, insights, and limitations

    Tom Stoops, Annick De Backer, Ivan Lobato, and Sandra Van Aert. Obtaining 3d atomic recon- structions from electron microscopy images using a bayesian genetic algorithm: Possibilities, insights, and limitations. Microscopy and Microanalysis, 31(1):ozae090, 2025

    2025

  13. [21]

    Auto-resolving the atomic structure at van der waals interfaces using a generative model

    Wenqiang Huang, Yucheng Jin, Zhemin Li, Lin Yao, Yun Chen, Zheng Luo, Shen Zhou, Jinguo Lin, Feng Liu, Zhifeng Gao, et al. Auto-resolving the atomic structure at van der waals interfaces using a generative model. Nature Communications, 16(1):2927, 2025

    2025

  14. [22]

    J. Li, M. Telychko, J. Yin, Y . Zhu, G. Li, S. Song, H. Yang, J. Li, J. Wu, J. Lu, and X. Wang. Machine vision automated chiral molecule detection and classification in molecular imaging. J Am Chem Soc, 143(27):10177–10188, 2021. ISSN 1520-5126 (Electronic) 0002-7863 (Linking)...

    2021

  15. [23]

    Deep learning analysis of defect and phase evolution during electron beam-induced transformations in ws2

    Artem Maksov, Ondrej Dyck, Kai Wang, Kai Xiao, David B Geohegan, Bobby G Sumpter, Rama K Vasudevan, Stephen Jesse, Sergei V Kalinin, and Maxim Ziatdinov. Deep learning analysis of defect and phase evolution during electron beam-induced transformations in ws2. npj Computational...

    2019

  16. [24]

    Machine-learning approach for quantified resolvability enhancement of low-dose stem data

    Laura Gambini, Tiarnan Mullarkey, Lewys Jones, and Stefano Sanvito. Machine-learning approach for quantified resolvability enhancement of low-dose stem data. Machine Learning: Science and Technology, 4(1), 2023. ISSN 2632-2153. doi: 10.1088/2632-2153/acbb52

    2023 doi

  17. [25]

    Deep learning-based noise filtering toward millisecond order imaging by using scanning trans- mission electron microscopy

    Shiro Ihara, Hikaru Saito, Mizumo Yoshinaga, Lavakumar Avala, and Mitsuhiro Murayama. Deep learning-based noise filtering toward millisecond order imaging by using scanning trans- mission electron microscopy. Scientific reports, 12(1):13462, 2022. ISSN 2045-2322

    2022

  18. [26]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023

    2023 arXiv

  19. [27]

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, et al. Qwen2. 5-vl technical report. arXiv preprint arXiv:2502.13923, 2025

    2025 arXiv

  20. [28]

    Atomai framework for deep learning analysis of image and spectroscopy data in electron and scanning probe microscopy

    Maxim Ziatdinov, Ayana Ghosh, Chun Yin Wong, and Sergei V Kalinin. Atomai framework for deep learning analysis of image and spectroscopy data in electron and scanning probe microscopy. Nature Machine Intelligence, 4(12):1101–1112, 2022

    2022

  21. [29]

    The computational 2d materials database: high-throughput modeling and discovery of atomically thin crystals

    Sten Haastrup, Mikkel Strange, Mohnish Pandey, Thorsten Deilmann, Per S Schmidt, Nicki F Hinsche, Morten N Gjerding, Daniele Torelli, Peter M Larsen, Anders C Riis-Jensen, et al. The computational 2d materials database: high-throughput modeling and discovery of atomically thin...

    2018

  22. [30]

    Commentary: The materials project: A materials genome approach to accelerating materials innovation

    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 materi...

    2013

  23. [31]

    Thiessen, Evan E

    Antanas Vaitkus, Andrius Merkys, Thomas Sander, Miguel Quirós, Paul A. Thiessen, Evan E. Bolton, and Saulius Gražulis. A workflow for deriving chemical entities from crystallographic data and its application to the Crystallography Open Database. Journal of Cheminformatics, 15(...

    2023 doi

  24. [32]

    abtem: Ab initio transmission electron microscopy image simulation

    Jacob Madsen and Toma Susi. abtem: Ab initio transmission electron microscopy image simulation. Microscopy and Microanalysis, 26(S2):448–450, 2020. 12

    2020

  25. [33]

    Chemvlm: Exploring the power of multimodal large language models in chemistry area

    Junxian Li, Di Zhang, Xunzhi Wang, Zeying Hao, Jingdi Lei, Qian Tan, Cai Zhou, Wei Liu, Yaotian Yang, Xinrui Xiong, et al. Chemvlm: Exploring the power of multimodal large language models in chemistry area. In Proceedings of the AAAI Conference on Artificial Intelligence, volu...

    2025

  26. [34]

    Chemcrow: Augmenting large-language models with chemistry tools, 2023

    Andres M Bran, Sam Cox, Oliver Schilter, Carlo Baldassari, Andrew D White, and Philippe Schwaller. Chemcrow: Augmenting large-language models with chemistry tools, 2023. URL https://arxiv.org/abs/2304.05376

    2023 arXiv

  27. [35]

    Chemagent: Self-updating library in large language models improves chemical reasoning, 2025

    Xiangru Tang, Tianyu Hu, Muyang Ye, Yanjun Shao, Xunjian Yin, Siru Ouyang, Wangchunshu Zhou, Pan Lu, Zhuosheng Zhang, Yilun Zhao, Arman Cohan, and Mark Gerstein. Chemagent: Self-updating library in large language models improves chemical reasoning, 2025. URL https://arxiv.org/...

    2025 arXiv

  28. [36]

    A universal graph deep learning interatomic potential for the periodic table

    Chi Chen and Shyue Ping Ong. A universal graph deep learning interatomic potential for the periodic table. Nature Computational Science, 2(11):718–728, 2022

    2022

  29. [37]

    Ilyes Batatia, Dávid Péter Kovács, Gregor N. C. Simm, Christoph Ortner, and Gábor Csányi. Mace: Higher order equivariant message passing neural networks for fast and accurate force fields, 2023. URL https://arxiv.org/abs/2206.07697

    2023 arXiv

  30. [38]

    Viewrefer: Grasp the multi-view knowledge for 3d visual grounding

    Zoey Guo, Yiwen Tang, Ray Zhang, Dong Wang, Zhigang Wang, Bin Zhao, and Xuelong Li. Viewrefer: Grasp the multi-view knowledge for 3d visual grounding. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 15372–15383, 2023

    2023

  31. [39]

    Any2point: Empowering any-modality large models for efficient 3d understanding

    Yiwen Tang, Ray Zhang, Jiaming Liu, Zoey Guo, Bin Zhao, Zhigang Wang, Peng Gao, Hong- sheng Li, Dong Wang, and Xuelong Li. Any2point: Empowering any-modality large models for efficient 3d understanding. In European Conference on Computer Vision, pages 456–473. Springer, 2024

    2024

  32. [40]

    Exploring the potential of encoder-free architectures in 3d lmms

    Yiwen Tang, Zoey Guo, Zhuhao Wang, Ray Zhang, Qizhi Chen, Junli Liu, Delin Qu, Zhigang Wang, Dong Wang, Xuelong Li, et al. Exploring the potential of encoder-free architectures in 3d lmms. arXiv preprint arXiv:2502.09620, 2025. 13

    2025

  33. [2024]

    URL https://arxiv.org/abs/2401.00096

    arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.