pith. sign in

arxiv: 2506.16255 · v2 · submitted 2025-06-19 · 🌌 astro-ph.IM · cs.AI

Category-based Galaxy Image Generation via Diffusion Models

Xingzhong Fan , Hongming Tang , Yue Zeng , M.B.N.Kouwenhoven , Guangquan Zeng This is my paper

Pith reviewed 2026-05-19 08:53 UTC · model grok-4.3

classification 🌌 astro-ph.IM cs.AI
keywords galaxy image generationdiffusion modelscategory embeddingsresidual attention blockastrophysical propertiesdata augmentationgalaxy simulations
0
0 comments X p. Extension

The pith

GalCatDiff generates galaxy images via diffusion models that match observed color and size distributions while using category embeddings to avoid training separate models per class.

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

The paper establishes that a diffusion-based framework can produce galaxy images by learning directly from observational data rather than relying on pre-set physical parameters in simulations. It introduces category embeddings so one model handles multiple galaxy types and adds an Astro-RAB block that mixes attention and convolution to keep global structure and local details aligned. If the approach holds, it offers a data-driven route to realistic galaxy catalogs that could supplement or replace parts of traditional semi-analytic modeling for large-scale surveys.

Core claim

GalCatDiff is the first astronomy-specific diffusion framework that folds both image features and astrophysical category information into the network through an enhanced U-Net and a Residual Attention Block; category embeddings enable class-conditioned generation without separate per-category trainings, and experiments show the outputs match real color and size distributions more closely than prior generative methods while appearing visually realistic.

What carries the argument

Category embeddings combined with the Astro-RAB (Residual Attention Block) that merges attention mechanisms and convolution operations inside the diffusion U-Net to enforce both global consistency and local feature fidelity.

If this is right

  • One trained model can produce galaxies from multiple categories without retraining separate networks for each class.
  • The generated images can serve directly as augmented training data for downstream galaxy classification tasks.
  • Galaxy simulations become less dependent on manual parameter tuning in semi-analytic models.
  • The framework scales to larger catalogs while preserving distribution consistency across samples.

Where Pith is reading between the lines

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

  • If physical consistency extends beyond color and size, the same conditioning approach could let researchers request galaxies with targeted properties such as specific star-formation rates.
  • The category-embedding technique might transfer to other image domains where classes carry physical meaning, such as generating spectra or light curves conditioned on object type.
  • Large synthetic catalogs produced this way could be used to stress-test cosmological inference pipelines before they are applied to real survey data.

Load-bearing premise

Matching color and size distributions plus visual inspection is enough to establish that the generated galaxies are physically consistent.

What would settle it

A test in which galaxies generated by the model are compared against an independent survey catalog on quantities such as redshift distribution, stellar mass, or morphological parameters that were never used as conditioning inputs.

Figures

Figures reproduced from arXiv: 2506.16255 by Guangquan Zeng, Hongming Tang, M.B.N.Kouwenhoven, Xingzhong Fan, Yue Zeng.

Figure 1
Figure 1. Figure 1: The schematic diagram of the diffusion model framework for category galaxy generation (GalCatDiff). In the forward process, a clean image X0 is progressively corrupted by adding Gaussian noise at each timestep t, eventually producing a pure noise image XT . The reverse process, guided by the enhanced U-Net estimating Pθ(Xt−1 | Xt), starts from the noise image XT and iteratively removes noise to reconstruct… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the Astro-RAB structure. Each block consists of a convolutional block followed by an Attention Fusion Unit and a skip connection, with this sequence iterated twice. The Attention Fusion Unit incorporates window attention, combined with convolutional layers, to preserve essential galaxy physics properties. The dataset used in this study was derived from the Galaxy10 DECaLS dataset1 , which conta… view at source ↗
Figure 3
Figure 3. Figure 3: A total of 311 images were excluded from the dataset to improve the quality and category accuracy of the generated results. The images were removed due to factors such as low quality, contamination, or the lack of a dominant central galaxy. For example, (a) images with significant color contamination, (b) images exhibiting excessive noise, (c) the central galaxy’s brightness is distorted, (d) strong light … view at source ↗
Figure 4
Figure 4. Figure 4: The images were generated by GalCatDiff after the 80k model training epoch of training on the Galaxy6 DECaLS dataset. For each of the six galaxy classes, ten images were generated as references. The g, r, and z channels were mapped to the r, g, and b channels. These generated galaxies maintain high image quality while preserving both inter-class and intra￾class diversity. Key distinguishing features betwee… view at source ↗
Figure 5
Figure 5. Figure 5: The histograms compare key features—g, r, and z band aperture magnitudes, color index g-r and r-z, and half-light radii—between the images generated by GalCatDiff (trained on the Galaxy6 DECaLS dataset) and the corresponding categories in the test set. The figure contains 42 distribution plots, organized into 6 rows and 7 columns, with each row representing a different feature and each column corresponding… view at source ↗
Figure 6
Figure 6. Figure 6: The parameters wattn and wconv are trainable weights from different Attention Fusion Unit layers of GalCatDiff. The output of each Attention Fusion Unit is determined by the weighted sum of the window attention and convolutional layers. The left, middle, and right plots illustrate the variations in wattn, wconv, and wattn/wconv across the down-sampling stages, the U-Net bottleneck layer, and the up-samplin… view at source ↗
read the original abstract

Conventional galaxy generation methods rely on semi-analytical models and hydrodynamic simulations, which are highly dependent on physical assumptions and parameter tuning. In contrast, data-driven generative models do not have explicit physical parameters pre-determined, and instead learn them efficiently from observational data, making them alternative solutions to galaxy generation. Among these, diffusion models outperform Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) in quality and diversity. Leveraging physical prior knowledge to these models can further enhance their capabilities. In this work, we present GalCatDiff, the first framework in astronomy to leverage both galaxy image features and astrophysical properties in the network design of diffusion models. GalCatDiff incorporates an enhanced U-Net and a novel block entitled Astro-RAB (Residual Attention Block), which dynamically combines attention mechanisms with convolution operations to ensure global consistency and local feature fidelity. Moreover, GalCatDiff uses category embeddings for class-specific galaxy generation, avoiding the high computational costs of training separate models for each category. Our experimental results demonstrate that GalCatDiff significantly outperforms existing methods in terms of the consistency of sample color and size distributions, and the generated galaxies are both visually realistic and physically consistent. This framework will enhance the reliability of galaxy simulations and can potentially serve as a data augmentor to support future galaxy classification algorithm development.

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 / 2 minor

Summary. The manuscript introduces GalCatDiff, a diffusion-model framework for generating galaxy images conditioned on astrophysical categories. It employs an enhanced U-Net architecture augmented by a novel Astro-RAB (Residual Attention Block) that fuses attention mechanisms with convolutional operations, together with category embeddings to enable class-specific generation without training separate models per category. The central empirical claim is that GalCatDiff produces samples whose color and size distributions are more consistent with real data than those from prior methods, while the images are visually realistic and physically consistent, offering a data-driven alternative to semi-analytical models and hydrodynamic simulations.

Significance. If the reported improvements are robustly demonstrated, the work would constitute a useful methodological contribution to astronomical image synthesis by embedding category-level physical priors directly into the generative architecture. The Astro-RAB design and single-model multi-category approach could reduce computational overhead for producing mock catalogs and support data-augmentation pipelines for downstream classification tasks. Credit is due for the explicit attempt to incorporate astrophysical properties into the network rather than treating generation as a purely unsupervised image task.

major comments (2)
  1. [Abstract] Abstract: the assertion that GalCatDiff 'significantly outperforms existing methods in terms of the consistency of sample color and size distributions' is presented without any quantitative metrics, baseline names, error bars, dataset sizes, or statistical tests. This absence prevents evaluation of the claimed improvement and directly undermines the headline result.
  2. [Abstract] Abstract and results: the claim that generated galaxies are 'physically consistent' rests solely on alignment of color and size marginal distributions plus visual inspection. No validation is reported against independent observables (e.g., Sérsic indices, morphological statistics, or stellar-mass functions) or against outputs from hydrodynamic simulations withheld from training. Because this inference is load-bearing for the central contribution, additional quantitative checks are required.
minor comments (2)
  1. The acronym Astro-RAB is introduced without an immediate parenthetical expansion on first use; spelling out 'Residual Attention Block' at its initial appearance would improve readability for readers outside the immediate subfield.
  2. Figure captions and axis labels should explicitly state the sample sizes and number of generated versus real galaxies used for the distribution comparisons to allow direct assessment of statistical power.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed review. The comments highlight important points about clarity in the abstract and the strength of evidence for physical consistency. We address each major comment below and indicate where revisions will be made to the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the assertion that GalCatDiff 'significantly outperforms existing methods in terms of the consistency of sample color and size distributions' is presented without any quantitative metrics, baseline names, error bars, dataset sizes, or statistical tests. This absence prevents evaluation of the claimed improvement and directly undermines the headline result.

    Authors: We agree that the abstract would benefit from greater specificity to allow immediate evaluation of the headline claim. The results section (Section 4) already contains the relevant quantitative comparisons, including baseline methods, dataset sizes from the observational catalog, error estimates from multiple sampling runs, and statistical tests (e.g., distribution similarity metrics) demonstrating outperformance. We will revise the abstract to incorporate concise references to these elements, such as the specific baselines, sample scale, and key statistical outcomes, while remaining within length constraints. revision: yes

  2. Referee: [Abstract] Abstract and results: the claim that generated galaxies are 'physically consistent' rests solely on alignment of color and size marginal distributions plus visual inspection. No validation is reported against independent observables (e.g., Sérsic indices, morphological statistics, or stellar-mass functions) or against outputs from hydrodynamic simulations withheld from training. Because this inference is load-bearing for the central contribution, additional quantitative checks are required.

    Authors: Color and size distributions are fundamental observables that encode substantial physical information (stellar populations, dust, and structural scaling relations), and their close match to real data, combined with expert visual assessment, forms the primary support for the consistency claim in the current study. We acknowledge that further checks against Sérsic indices or morphological statistics would strengthen the argument. We will add a dedicated paragraph in the results and discussion sections that reports any available morphological comparisons from the dataset and explicitly discusses the limitations of the current validation. Comparisons to hydrodynamic simulations withheld from training fall outside the scope of this observational data-driven work; we will note this as a limitation and suggest it as a direction for future research. revision: partial

Circularity Check

0 steps flagged

No significant circularity; results are empirical outputs of a trained model

full rationale

The paper presents a data-driven diffusion model (GalCatDiff) that learns galaxy image generation from observational data without explicit physical parameters or algebraic derivations. Central claims rest on empirical comparisons of color/size distributions and visual realism from the trained network outputs. No load-bearing steps reduce by construction to inputs via self-definition, fitted parameters renamed as predictions, or self-citation chains. The model architecture (enhanced U-Net with Astro-RAB) and category embeddings are design choices evaluated experimentally, not forced equivalences. This is a standard self-contained ML training/evaluation setup with no circular reduction in the reported results.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The framework rests on the standard diffusion-model assumption that observational images contain sufficient information to learn realistic distributions, plus the design choice that category embeddings plus a single shared network can replace separate per-class models. The Astro-RAB block is introduced as a novel architectural component without external validation cited in the abstract.

axioms (1)
  • domain assumption Data-driven generative models can learn physical regularities directly from observational images without explicit pre-determined physical parameters.
    Stated in the abstract as the core contrast with semi-analytical and hydrodynamic methods.
invented entities (1)
  • Astro-RAB (Residual Attention Block) no independent evidence
    purpose: Dynamically combine attention mechanisms with convolution operations to ensure global consistency and local feature fidelity in galaxy images.
    Novel block introduced in the GalCatDiff network design; no independent evidence outside the paper is provided in the abstract.

pith-pipeline@v0.9.0 · 5771 in / 1404 out tokens · 46500 ms · 2026-05-19T08:53:03.904095+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

106 extracted references · 106 canonical work pages · 6 internal anchors

  1. [1]

    2018, Monthly Notices of the Royal Astronomical Society, 477, 894

    Vaghmare, K. 2018, Monthly Notices of the Royal Astronomical Society, 477, 894

    work page 2018

  2. [2]

    2024, Nature, 1

    Abramson, J., Adler, J., Dunger, J., et al. 2024, Nature, 1

    work page 2024

  3. [3]

    2020, Astronomy & Astrophysics, 643, A177

    Angora, G., Rosati, P., Brescia, M., et al. 2020, Astronomy & Astrophysics, 643, A177

    work page 2020

  4. [4]

    2017, The Astrophysical Journal Supplement Series, 230, 20

    Aniyan, A., & Thorat, K. 2017, The Astrophysical Journal Supplement Series, 230, 20

    work page 2017

  5. [5]

    Arcelin, B., Doux, C., Aubourg, E., Roucelle, C., & Collaboration), L. D. E. S. 2021, Monthly Notices of the Royal Astronomical Society, 500, 531

    work page 2021

  6. [6]

    2024, Astronomy & Astrophysics, 683, A181

    Baes, M., Gebek, A., Trˇ cka, A., et al. 2024, Astronomy & Astrophysics, 683, A181

    work page 2024

  7. [7]

    Neural Machine Translation by Jointly Learning to Align and Translate

    Bahdanau, D. 2014, arXiv preprint arXiv:1409.0473

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  8. [8]

    and Glazebrook, Karl and Brinkmann, Jon and Ivezić, Željko and Lupton, Robert H

    Baldry, I. K., Glazebrook, K., Brinkmann, J., et al. 2004, ApJ, 600, 681, doi: 10.1086/380092

    work page doi:10.1086/380092 2004

  9. [9]

    A., Pedrosa, S

    Bignone, L. A., Pedrosa, S. E., Trayford, J. W., Tissera, P. B., & Pellizza, L. J. 2020, Monthly Notices of the Royal Astronomical Society, 491, 3624

    work page 2020

  10. [10]

    2020, Monthly Notices of the Royal Astronomical Society, 491, 2481

    Boucaud, A., Huertas-Company, M., Heneka, C., et al. 2020, Monthly Notices of the Royal Astronomical Society, 491, 2481

    work page 2020

  11. [11]

    Buta, R. J. 2013, in Secular Evolution of Galaxies, ed. J. Falc´ on-Barroso & J. H. Knapen, 155, doi: 10.48550/arXiv.1304.3529

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1304.3529 2013

  12. [12]

    J., Sheth, K., Athanassoula, E., et al

    Buta, R. J., Sheth, K., Athanassoula, E., et al. 2015, ApJS, 217, 32, doi: 10.1088/0067-0049/217/2/32

    work page doi:10.1088/0067-0049/217/2/32 2015

  13. [13]

    2016, ARA&A, 54, 597, doi: 10.1146/annurev-astro-082214-122432

    Cappellari, M. 2016, ARA&A, 54, 597, doi: 10.1146/annurev-astro-082214-122432

    work page doi:10.1146/annurev-astro-082214-122432 2016

  14. [14]

    Mode Regularized Generative Adversarial Networks

    Che, T., Li, Y., Jacob, A. P., Bengio, Y., & Li, W. 2016, arXiv e-prints, arXiv:1612.02136, doi: 10.48550/arXiv.1612.02136

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1612.02136 2016

  15. [15]

    J., Arag´ on-Salamanca, A., et al

    Cheng, T.-Y., Conselice, C. J., Arag´ on-Salamanca, A., et al. 2020, Monthly Notices of the Royal Astronomical Society, 493, 4209 ´Ciprijanovi´ c, A., Kafkes, D., Downey, K., et al. 2021, Monthly Notices of the Royal Astronomical Society, 506, 677

    work page 2020

  16. [16]

    R., Debattista, V

    Cole, D. R., Debattista, V. P., Erwin, P., Earp, S. W. F., & Roˇ skar, R. 2014, MNRAS, 445, 3352, doi: 10.1093/mnras/stu1985

    work page doi:10.1093/mnras/stu1985 2014

  17. [17]

    A unified model for the evolution of galaxies and quasars , volume =

    Cole, S., Lacey, C. G., Baugh, C. M., & Frenk, C. S. 2000, MNRAS, 319, 168, doi: 10.1046/j.1365-8711.2000.03879.x

    work page doi:10.1046/j.1365-8711.2000.03879.x 2000

  18. [18]

    Conselice, C. J. 2014, ARA&A, 52, 291, doi: 10.1146/annurev-astro-081913-040037

    work page doi:10.1146/annurev-astro-081913-040037 2014

  19. [19]

    A., Schaye, J., Clauwens, B., et al

    Correa, C. A., Schaye, J., Clauwens, B., et al. 2017, Monthly Notices of the Royal Astronomical Society: Letters, 472, L45

    work page 2017

  20. [20]

    A., & van de Voort, F

    Crain, R. A., & van de Voort, F. 2023, Annual Review of Astronomy and Astrophysics, 61, 473

    work page 2023

  21. [21]

    T., & Shah, M

    Croitoru, F.-A., Hondru, V., Ionescu, R. T., & Shah, M. 2023, IEEE Transactions on Pattern Analysis and Machine Intelligence, 45, 10850

    work page 2023

  22. [22]

    , keywords =

    Croton, D. J., Springel, V., White, S. D. M., et al. 2006, MNRAS, 365, 11, doi: 10.1111/j.1365-2966.2005.09675.x

    work page doi:10.1111/j.1365-2966.2005.09675.x 2006

  23. [23]

    V., & Tan, M

    Dai, Z., Liu, H., Le, Q. V., & Tan, M. 2021, Advances in neural information processing systems, 34, 3965

    work page 2021

  24. [24]

    J., Lang, D., et al

    Dey, A., Schlegel, D. J., Lang, D., et al. 2019, The Astronomical Journal, 157, 168

    work page 2019

  25. [25]

    2021, Advances in neural information processing systems, 34, 8780

    Dhariwal, P., & Nichol, A. 2021, Advances in neural information processing systems, 34, 8780

    work page 2021

  26. [26]

    W., & Dambre, J

    Dieleman, S., Willett, K. W., & Dambre, J. 2015, Monthly notices of the royal astronomical society, 450, 1441

    work page 2015

  27. [27]

    , month =

    Dressler, A. 1980, ApJ, 236, 351, doi: 10.1086/157753

    work page doi:10.1086/157753 1980

  28. [28]

    Pal, C. 2016, in International workshop on deep learning in medical image analysis, international workshop on large-scale annotation of biomedical data and expert label synthesis, Springer, 179–187

    work page 2016

  29. [29]

    2018, Monthly Notices of the Royal Astronomical Society, 473, 1930 Euclid Collaboration, Bretonni` ere, H., Huertas-Company, M., et al

    El-Badry, K., Quataert, E., Wetzel, A., et al. 2018, Monthly Notices of the Royal Astronomical Society, 473, 1930 Euclid Collaboration, Bretonni` ere, H., Huertas-Company, M., et al. 2022, A&A, 657, A90, doi: 10.1051/0004-6361/202141393 Euclid Collaboration, Bretonni` ere, H., Kuchner, U., et al. 2023, A&A, 671, A102, doi: 10.1051/0004-6361/202245042 Eucl...

    work page doi:10.1051/0004-6361/202141393 2018

  30. [30]

    J., Kuchner, U., & Tohill, C.-B

    Ferreira, L., Conselice, C. J., Kuchner, U., & Tohill, C.-B. 2022, The Astrophysical Journal, 931, 34

    work page 2022

  31. [31]

    2019, Monthly Notices of the Royal Astronomical Society, 485, 3203

    Fussell, L., & Moews, B. 2019, Monthly Notices of the Royal Astronomical Society, 485, 3203

    work page 2019

  32. [32]

    2024, Advances in Neural Information Processing Systems, 36

    Gao, Z., Shi, X., Han, B., et al. 2024, Advances in Neural Information Processing Systems, 36

    work page 2024

  33. [33]

    Ghiasi, G., Lin, T.-Y., & Le, Q. V. 2018, Advances in neural information processing systems, 31

    work page 2018

  34. [34]

    G., Baugh, C

    Gonzalez-Perez, V., Lacey, C. G., Baugh, C. M., et al. 2014, MNRAS, 439, 264, doi: 10.1093/mnras/stt2410

    work page doi:10.1093/mnras/stt2410 2014

  35. [35]

    2014, Advances in neural information processing systems, 27

    Goodfellow, I., Pouget-Abadie, J., Mirza, M., et al. 2014, Advances in neural information processing systems, 27

    work page 2014

  36. [36]

    Galactic Centre region, longitudes 345 to 6

    Guo, Q., White, S., Boylan-Kolchin, M., et al. 2011, MNRAS, 413, 101, doi: 10.1111/j.1365-2966.2010.18114.x

    work page doi:10.1111/j.1365-2966.2010.18114.x 2011

  37. [37]

    2016, in Proceedings of the IEEE conference on computer vision and pattern recognition, 770–778 17

    He, K., Zhang, X., Ren, S., & Sun, J. 2016, in Proceedings of the IEEE conference on computer vision and pattern recognition, 770–778 17

    work page 2016

  38. [38]

    2017, Advances in neural information processing systems, 30

    Hochreiter, S. 2017, Advances in neural information processing systems, 30

    work page 2017

  39. [39]

    2020, Advances in neural information processing systems, 33, 6840

    Ho, J., Jain, A., & Abbeel, P. 2020, Advances in neural information processing systems, 33, 6840

    work page 2020

  40. [40]

    C., Filippenko, A

    Ho, L. C., Filippenko, A. V., & Sargent, W. L. W. 1997, ApJ, 487, 591, doi: 10.1086/304643

    work page doi:10.1086/304643 1997

  41. [41]

    2021, Astronomy & Astrophysics, 653, A76

    Holdship, J., Viti, S., Haworth, T., & Ilee, J. 2021, Astronomy & Astrophysics, 653, A76

    work page 2021

  42. [42]

    2022, Monthly Notices of the Royal Astronomical Society, 515, 652

    Rodriguez-Gomez, V., & Thuerey, N. 2022, Monthly Notices of the Royal Astronomical Society, 515, 652

    work page 2022

  43. [43]

    2024a, A&A, 691, A125, doi: 10.1051/0004-6361/202451732

    Hu, J., Cui, Q., Wang, L., Pei, W., & Ge, J. 2024a, A&A, 691, A125, doi: 10.1051/0004-6361/202451732

    work page doi:10.1051/0004-6361/202451732

  44. [44]

    2024b, MNRAS, 529, 4565, doi: 10.1093/mnras/stae827

    Hu, J., Wang, L., Ge, J., Zhu, K., & Zeng, G. 2024b, MNRAS, 529, 4565, doi: 10.1093/mnras/stae827

    work page doi:10.1093/mnras/stae827

  45. [45]

    Hubble, E. P. 1926, ApJ, 64, 321, doi: 10.1086/143018

    work page doi:10.1086/143018 1926

  46. [46]

    A., Martin, G., Kaviraj, S., et al

    Jackson, R. A., Martin, G., Kaviraj, S., et al. 2020, Monthly Notices of the Royal Astronomical Society, 494, 5568

    work page 2020

  47. [47]

    Joseph, V. R. 2022, Statistical Analysis and Data Mining: The ASA Data Science Journal, 15, 531

    work page 2022

  48. [48]

    2020, in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 8110–8119

    Karras, T., Laine, S., Aittala, M., et al. 2020, in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 8110–8119

    work page 2020

  49. [49]

    Kauffmann, G., White, S. D. M., & Guiderdoni, B. 1993, MNRAS, 264, 201, doi: 10.1093/mnras/264.1.201

    work page doi:10.1093/mnras/264.1.201 1993

  50. [50]

    S., Driver, S

    Kelvin, L. S., Driver, S. P., Robotham, A. S. G., et al. 2014, MNRAS, 439, 1245, doi: 10.1093/mnras/stt2391

    work page doi:10.1093/mnras/stt2391 2014

  51. [51]

    Kingma, D. P. 2013, arXiv preprint arXiv:1312.6114 —. 2014, arXiv preprint arXiv:1412.6980

    work page internal anchor Pith review Pith/arXiv arXiv 2013

  52. [52]

    2014, Advances in neural information processing systems, 27

    Welling, M. 2014, Advances in neural information processing systems, 27

    work page 2014

  53. [53]

    2011 , month =

    Kormendy, J., & Bender, R. 2012, ApJS, 198, 2, doi: 10.1088/0067-0049/198/1/2

    work page doi:10.1088/0067-0049/198/1/2 2012

  54. [54]

    Krizhevsky, A., Sutskever, I., & Hinton, G. E. 2012, Advances in neural information processing systems, 25

    work page 2012

  55. [55]

    2021, Monthly Notices of the Royal Astronomical Society, 504, 5543

    Lanusse, F., Mandelbaum, R., Ravanbakhsh, S., et al. 2021, Monthly Notices of the Royal Astronomical Society, 504, 5543

    work page 2021

  56. [56]

    G., et al

    Lee, G.-H., Woo, J.-H., Lee, M. G., et al. 2012, ApJ, 750, 141, doi: 10.1088/0004-637X/750/2/141

    work page doi:10.1088/0004-637x/750/2/141 2012

  57. [57]

    Li, Z., Shen, J., Gerhard, O., & Clarke, J. P. 2022, ApJ, 925, 71, doi: 10.3847/1538-4357/ac3823

    work page doi:10.3847/1538-4357/ac3823 2022

  58. [58]

    2011, Monthly Notices of the Royal Astronomical Society, 410, 166

    Lintott, C., Schawinski, K., Bamford, S., et al. 2011, Monthly Notices of the Royal Astronomical Society, 410, 166

    work page 2011

  59. [59]

    2021, in Proceedings of the IEEE/CVF international conference on computer vision, 10012–10022

    Liu, Z., Lin, Y., Cao, Y., et al. 2021, in Proceedings of the IEEE/CVF international conference on computer vision, 10012–10022

    work page 2021

  60. [60]

    2016, Advances in neural information processing systems, 29

    Mao, X., Shen, C., & Yang, Y.-B. 2016, Advances in neural information processing systems, 29

    work page 2016

  61. [61]

    2018, Monthly Notices of the Royal Astronomical Society, 480, 2266

    Pichon, C. 2018, Monthly Notices of the Royal Astronomical Society, 480, 2266

    work page 2018

  62. [62]

    Attention U-Net: Learning Where to Look for the Pancreas

    Oktay, O. 2018, arXiv preprint arXiv:1804.03999

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  63. [63]

    2022, in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 815–825

    Pan, X., Ge, C., Lu, R., et al. 2022, in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 815–825

    work page 2022

  64. [64]

    2019, Advances in neural information processing systems, 32

    Paszke, A., Gross, S., Massa, F., et al. 2019, Advances in neural information processing systems, 32

    work page 2019

  65. [65]

    K., Bekki, K., et al

    Pfeffer, J., Cavanagh, M. K., Bekki, K., et al. 2023, Monthly Notices of the Royal Astronomical Society, 518, 5260

    work page 2023

  66. [66]

    2023, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 18434–18443

    Qin, Y., Zheng, H., Yao, J., Zhou, M., & Zhang, Y. 2023, in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 18434–18443

    work page 2023

  67. [67]

    A., & Baso, C

    Ramos, A. A., & Baso, C. D. 2019, Astronomy & Astrophysics, 626, A102

    work page 2019

  68. [68]

    2017, in Proceedings of the AAAI Conference on Artificial Intelligence, Vol

    Ravanbakhsh, S., Lanusse, F., Mandelbaum, R., Schneider, J., & Poczos, B. 2017, in Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 31

    work page 2017

  69. [69]

    W., Parul, H., & Gleyzer, S

    Reddy, P., Toomey, M. W., Parul, H., & Gleyzer, S. 2024, Machine Learning: Science and Technology, 5, 035076

    work page 2024

  70. [70]

    W., & Teuben, P

    Regan, M. W., & Teuben, P. J. 2004, ApJ, 600, 595, doi: 10.1086/380116

    work page doi:10.1086/380116 2004

  71. [71]

    W., Huntley, J

    Roberts, Jr., W. W., Huntley, J. M., & van Albada, G. D. 1979, ApJ, 233, 67, doi: 10.1086/157367

    work page doi:10.1086/157367 1979

  72. [72]

    F., Lotz, J

    Rodriguez-Gomez, V., Snyder, G. F., Lotz, J. M., et al. 2019, Monthly Notices of the Royal Astronomical Society, 483, 4140

    work page 2019

  73. [73]

    Ronneberger, O., Fischer, P., & Brox, T. 2015, in Medical image computing and computer-assisted intervention–MICCAI 2015: 18th international conference, Munich, Germany, October 5-9, 2015, proceedings, part III 18, Springer, 234–241

    work page 2015

  74. [74]

    2016, Advances in neural information processing systems, 29 S´ anchez, H

    Salimans, T., Goodfellow, I., Zaremba, W., et al. 2016, Advances in neural information processing systems, 29 S´ anchez, H. D., Martin, G., Damjanov, I., et al. 2023, Monthly Notices of the Royal Astronomical Society, 521, 3861

    work page 2016

  75. [75]

    1961, The Hubble Atlas of Galaxies (Carnegie Institute of Washington, Washington)

    Sandage, A. 1961, The Hubble Atlas of Galaxies (Carnegie Institute of Washington, Washington)

    work page 1961

  76. [76]

    Shlosman, I., Frank, J., & Begelman, M. C. 1989, Nature, 338, 45, doi: 10.1038/338045a0

    work page doi:10.1038/338045a0 1989

  77. [77]

    P., & Devabhaktuni, V

    Siddique, N., Paheding, S., Elkin, C. P., & Devabhaktuni, V. 2021, IEEE access, 9, 82031

    work page 2021

  78. [78]

    Very Deep Convolutional Networks for Large-Scale Image Recognition

    Simonyan, K., & Zisserman, A. 2014, arXiv preprint arXiv:1409.1556

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  79. [79]

    J., & Geach, J

    Smith, M. J., & Geach, J. E. 2019, Monthly Notices of the Royal Astronomical Society, 490, 4985 18

    work page 2019

  80. [80]

    J., Geach, J

    Smith, M. J., Geach, J. E., Jackson, R. A., et al. 2022, Monthly Notices of the Royal Astronomical Society, 511, 1808

    work page 2022

Showing first 80 references.