Pith. sign in

REVIEW 3 major objections 8 minor 2 cited by

Radio Halo Detection in MWA Data using Deep Neural Networks and Generative Data Augmentation

T0 review · 3 major / 8 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read This paper claims that synthetic radio halo images produced by a denoising diffusion probabilistic model can train a neural classifier to find diffuse halos in MWA/GLEAM data without X-ray or SZ preselection, reaching about 96% validation…

desk verdict A plausible small-data ML recipe with an uncalibrated detection claim: the internal classifier comparison is fine, but the external halo detection rates need a false-positive baseline before they mean anything. read the letter →

arxiv 2411.15559 v1 pith:XLZRNPEW submitted 2024-11-23 astro-ph.GA

Ashutosh K. Mishra , Emma Tolley , Shreyam Parth Krishna , Jean-Paul Kneib This is my paper · ORCID
classification astro-ph.GA
keywords radiohalosgalaxyclustersdiffuseemissiondeeplearningdenoisingdiffusionprobabilisticmodelsgenerativedataaugmentationMurchisonWidefieldArrayGLEAMsurvey
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

Radio halos in galaxy clusters are usually found by first selecting clusters in X-ray or Sunyaev-Zeldovich surveys, which biases which halos are ever seen. This paper asks whether a neural classifier can detect halos directly from Murchison Widefield Array images, and whether synthetic halo images generated by a denoising diffusion probabilistic model (DDPM) can substitute for the very small number of real labeled examples. The authors train a binary halo-versus-non-halo classifier on multi-frequency 64x64 cutouts from the GLEAM survey, augmenting 18 halo sources with DDPM-generated images. The diffusion-supported attention classifier reaches an average validation accuracy of 95.93%, rediscovers 9 of 12 MGCLS halos and 5 of 8 PSZ2 halos inside GLEAM data, and flags 11 new COSMOS candidates. If the detection rates hold, generative augmentation offers a route to unbiased, survey-scale halo searches ahead of the next generation of radio telescopes.

What carries the argument

The load-bearing mechanism is generative data augmentation with a Denoising Diffusion Probabilistic Model (DDPM): the model corrupts images by adding Gaussian noise in a forward Markov chain and learns to reverse the process with a U-Net, producing synthetic (64, 64, 4) multi-frequency halo cutouts that are added to the training set. The classifier is a convolutional network augmented with a multi-head self-attention block, which captures correlations between distant pixels with fewer trainable parameters than a plain CNN. The paper uses the FID and KID metrics to show DDPM-generated halos are closer to real images than WGAN-generated ones, and shows that adding diffusion samples improves every classifier variant, with the attention variant benefiting most.

What would settle it

Run the trained classifier on hundreds of random blank sky patches and known non-halo radio galaxies; if the fraction scoring above 0.95 approaches the claimed detection rates (75% and 63%), then the threshold is not specific to halos and the rediscovery numbers would not demonstrate halo detection.

Watch

Extended reading notes

Core claim

The paper establishes that a small real sample of MWA halo images can be supplemented with DDPM-generated multi-frequency images to train a classifier that distinguishes halos from non-halos, and that this classifier generalizes to sources it was never trained on. The best variant, a convolutional network with a multi-head attention block, achieves 95.93% average validation accuracy and an F1 score of 0.965 over ten runs, using 36 clusters for training and 10 for testing with no hyperparameter tuning. Applied to external catalogues, it recovers 9/12 halos from the MeerKAT Galaxy Cluster Legacy Survey and 5/8 halos from the Planck SZ Catalogue 2 within GLEAM, and it proposes 11 XMM-Chandra-detected COSMOS clusters as new halo, minihalo, or candidate-halo sources.

Load-bearing premise

The load-bearing premise is that a classifier score of at least 0.95 marks a real halo rather than a compact source, artifact, or unrelated diffuse feature; the paper does not measure how often the same score occurs on empty or non-halo positions.

Editorial extensions

If this is right

  • A classifier trained on low-resolution MWA data can screen higher-resolution surveys for diffuse halo candidates without requiring X-ray or SZ preselection.
  • The 11 COSMOS candidates become a concrete follow-up target list; confirming any of them would show the tool discovers new objects rather than merely re-finding known ones.
  • Generative augmentation with diffusion models can be reused for other radio-source classification problems where labeled samples number in the tens.
  • The validation accuracy is measured on a deliberately simple binary task; extending the same pipeline to more diverse non-halo classes and multi-survey images is the paper's own stated next step.
  • The generated halo images themselves are released as a labeled dataset, giving the community a resource for further classifier development.

Reading between the lines

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

  • Beyond the paper, the detection rates at the 0.95 threshold are upper envelopes until the false-positive rate on empty fields and compact non-halo sources is measured, since the paper does not report that calibration.
  • Beyond the paper, the fixed ratio of generated to real training samples (about 0.5) is left untuned, so both higher and lower synthetic-to-real ratios could change the accuracy and the generalization to faint halos.
  • Beyond the paper, applying the same classifier to native-resolution MGCLS or LOFAR images would test whether diffusion augmentation transfers across resolution regimes, which the current GLEAM-only training does not directly establish.
  • Beyond the paper, moving classification from the image plane to the u-v plane, as the conclusion suggests, would make the method independent of imaging artifacts and could enable blind all-sky searches.
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

3 major / 8 minor

Summary. The paper trains binary classifiers (a LeNet-style CNN, a dense network, and a CNN with multi-head attention) to detect radio halos in 64x64 4-frequency cutouts from MWA/GLEAM observations, using a small labeled sample of 23 halo and 23 non-halo sources. To overcome the paucity of real examples, the authors compare WGAN- and DDPM-generated halo images as data augmentation, report that DDPM images are more realistic by FID/KID, and find that a diffusion-augmented attention classifier achieves 95.93% accuracy on a held-out test partition. The classifier is then applied to catalog-selected cluster centers, rediscovering 9/12 MGCLS and 5/8 PSZ2 halos in GLEAM and flagging 11 XMM-Chandra clusters in the COSMOS field as candidate halos. The paper concludes that generative augmentation can support ML-based detection of diffuse emission.

Significance. If the results hold, this is a useful methodological demonstration: DDPM-based generative augmentation appears to help a simple classifier on a very small astronomical image dataset, and the problem (low-frequency halo detection without X-ray/SZ preselection) is timely. The systematic comparison of WGAN versus DDPM, the ablation across three classifier architectures, and the authors' explicit caveat that their detections indicate central diffuse emission rather than confirmed halos (Section 5.5.3) are strengths. The reproducible-code commitment is also a strength, though the repository link is not actually given in this version. However, the externally-facing claims -- the 9/12 and 5/8 rediscovery rates and the COSMOS candidate list -- are uncalibrated: no negative-control false-positive measurement is presented for the score>=0.95 threshold, and the PSZ2 denominator is adjusted post hoc. These issues currently prevent the paper from supporting the 'unbiased detection' language in the abstract and from establishing the specific detection fractions as reliable.

major comments (3)
  1. [5.4] The detection threshold of classification score >= 0.95 is introduced 'to minimize false positives', but no false-positive rate is ever measured. The classifier is applied only to catalog-selected cluster centers (MGCLS, PSZ2, XMM-Chandra), and the paper's own Section 5.5.3 states that the detections should be seen as 'identification of a central source of diffuse emission in the cluster' rather than confirmed halos. Without a negative-control sample -- random sky positions, cluster centers with no known diffuse halo, or artifact-rich fields -- the reported 9/12 and 5/8 rediscovery rates and the 11 COSMOS candidates do not demonstrate that the threshold is specific to halos rather than to compact or central diffuse emission at MWA resolution. I recommend adding null-field and non-halo cluster-center tests and reporting precision-recall as a function of the score threshold.
  2. [5.4.1] The denominator for the PSZ2 detection rate is chosen post hoc. The paper analyzes 17 PSZ2 halos in GLEAM, then excludes 9 (3 truncated, 4 noisy, 2 small) and reports 5/8. If all 17 are counted, the detection rate is 5/17 approximately 29%. The exclusion criteria ('truncated', 'high noise', 'smaller-sized') are not pre-registered and may be correlated with detectability, so the reported 63% detection rate is not robust. Please report the full-sample result and present the exclusions as a sensitivity analysis with quantitative definitions for each exclusion category.
  3. [5.3.2, Table 3] The held-out test set contains only 5 halo and 5 non-halo real sources, from which 1280 and 960 images are created by augmentation. The 10-run mean and standard deviation of accuracy and F1 therefore reflect repeated, highly correlated views of the same few sources, not generalization to new clusters. The paper acknowledges in Section 6 that k-fold validation is needed, but as it stands the headline accuracy of 95.93% is not supported as a generalizable performance metric. Please add source-level cross-validation (e.g., leave-one-cluster-out) and report per-source predictions.
minor comments (8)
  1. [Abstract, Table 5] The abstract reports 'best average validation accuracy of 95.93%', but Table 5 and Section 5.3.2 describe metrics calculated on the held-out test set; please use consistent terminology. Also, 'without further hyperparameter tuning' should be qualified by the early stopping and checkpoint selection based on validation accuracy.
  2. [Table 3] With 8 rotations x 2 flips x 10 brightness scalings, 18 training halo sources should yield 2880 images, not the listed 5440; similarly the other counts do not match the stated augmentation factors. Please clarify whether multiple frequency channels are treated as separate images or whether side cutouts are included, and correct the table so the counts are reproducible.
  3. [Figure 10 caption] The caption states 'Class 0 and Class 1, represent halos and non-halos respectively', which is reversed relative to Section 4.1 where class 1 is halos and class 0 is non-halos; the caption should be corrected.
  4. [5.4.2, Table 8] Several COSMOS candidates are within approximately 1-2 arcminutes of each other (e.g., CXOC J100024.9+023956, CXOC J100022.7+023801, CXOC J100024.6+023748), suggesting duplicate detections of the same system; the count of '11 potential new halos' should be de-duplicated with a matching radius.
  5. [Data Availability] The GitHub repository is listed as '(github-radio-halo-classification)' rather than a working URL; please provide the actual repository address and, if available, a dataset DOI.
  6. [2.2] The notation theta in [0,9] intersect Z is undefined and is not used later in the paper; either define it or remove it.
  7. [Abstract, Section 1] The abstract's 'capable of unbiased detection of diffuse emission' conflicts with Section 1's statement that the tool is 'designed for halo searches specifically in MWA data, using cataloged sources as references', since the external application is only to catalog-selected cluster centers. Please remove or limit the 'unbiased' claim to the labeling methodology rather than the survey strategy.
  8. [5.2] The FID/KID values for multi-frequency generated images are computed against the training set 'as is the standard practice', so they do not measure how well generated images match held-out halos; please state this limitation explicitly or evaluate on a held-out set.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: external benchmark sources (MGCLS, PSZ2, XMM-Chandra) are absent from training, so the detection claims are not forced by construction; the main caveats are calibration and test-set pseudoreplication, not input-output equivalence.

full rationale

No load-bearing step in the derivation reduces to its own inputs. The WGAN and DDPM generators are trained on real GLEAM halo cutouts, and the classifier is trained on classic plus generative augmentations of 18 halo and 18 non-halo sources (Table 3). The external evaluation uses MGCLS, PSZ2, and XMM-Chandra cluster images that are not in the training set, so the reported 9/12, 5/8, and 11-candidate results are not equal by construction to the training labels. The Section 5.4 score-threshold of 0.95 is an analyst-chosen operating point and is not fitted to the external catalogs, so it is a calibration concern rather than a fitted input renamed as a prediction. The post-hoc exclusion of 9 of 17 PSZ2 sources (3 truncated, 4 high-noise, 2 small) changes the denominator and weakens the 5/8 figure as an unbiased detection rate, but this is a sample-selection limitation, not circularity. The 95.93% validation accuracy is computed on augmented copies of the same 5 halo and 5 non-halo test sources, creating source-level pseudoreplication and overstating the independence of the 10-run benchmark; the paper itself states in Section 6 that k-fold validation is left for future work. Section 5.5.3 explicitly downgrades the external detections to identification of central diffuse emission rather than confirmed halos, which limits the astrophysical claim but does not make the pipeline circular. No self-citation chain, imported uniqueness theorem, or ansatz smuggled via citation is present.

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

The paper introduces no new physical entities. It relies on several domain assumptions about label quality, generated-image realism, survey representativeness, and metric validity, plus a few hand-chosen preprocessing and threshold parameters that directly affect the reported detection rates.

free parameters (4)
  • detection score threshold = 0.95
    Chosen by hand to 'minimize false positives' in Section 5.4 without measuring the false-positive rate; directly determines the 9/12 and 5/8 detection rates.
  • generated-to-real augmentation ratio Gamma = text says about 0.5; counts imply about 1.07
    Ratio of diffusion-generated to real training samples; the authors note in Section 5.5.2 that it remains a hyperparameter for future tuning, so the best value is not established.
  • 3-sigma background clipping threshold = 3
    Pixel values below 3 times the local RMS noise are set to zero during preprocessing in Section 3; changing this changes what the classifier sees.
  • 64 by 64 pixel cutout size = 64
    Chosen so halos fit inside the cutout while keeping image generation tractable, as described in Section 3; a different cutout changes the learned morphology.
assumptions (4)
  • domain assumption The literature labels used as ground truth (H, cH, mH, cmH vs non-halo) are correct, and candidate labels represent true diffuse emission.
    The positive training class is 23 sources, over half of which are candidates (Section 3, Table 1); if candidate labels are wrong, the classifier learns 'central diffuse emission' rather than confirmed halos, which the authors concede in Section 5.5.3.
  • domain assumption Diffusion-generated halo images are realistic enough that training on them improves generalization instead of teaching generator artifacts.
    The paper relies on FID/KID scores and visual inspection (Section 5.1) rather than a task-specific test of whether generated images carry the physical features needed for detection.
  • domain assumption MWA/GLEAM images after preprocessing are representative of the cluster fields the classifier is applied to.
    The classifier is trained on GLEAM cutouts and applied to GLEAM cutouts, but external sources vary in resolution, noise, and halo size; the authors exclude small and noisy PSZ2 sources as unsuitable in Section 5.4.1.
  • domain assumption InceptionV3 features trained on ImageNet provide a valid quality metric for single-channel radio images.
    FID and KID use an ImageNet-trained InceptionV3 network (Section 2.3); there is no evidence this feature space is calibrated for radio flux maps.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Radio Halo Detection in MWA Data using Deep Neural Networks and Generative Data Augmentation." pith.science (2026). https://pith.science/paper/XLZRNPEW

@misc{pith2026241115559,
  author       = {Pith},
  title        = {Pith review of: Radio Halo Detection in MWA Data using Deep Neural Networks and Generative Data Augmentation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XLZRNPEW}},
  note         = {Machine review of arXiv:2411.15559}
}
read the original abstract

Detecting diffuse radio emission, such as from halos, in galaxy clusters is crucial for understanding large-scale structure formation in the universe. Traditional methods, which rely on X-ray and Sunyaev-Zeldovich (SZ) cluster pre-selection, introduce biases that limit our understanding of the full population of diffuse radio sources. In this work, we provide a possible resolution for this astrophysical tension by developing a machine learning (ML) framework capable of unbiased detection of diffuse emission, using a limited real dataset like those from the Murchison Widefield Array (MWA). We generate for the first time radio halo images using Wasserstein Generative Adversarial Networks (WGANs) and Denoising Diffusion Probabilistic Models (DDPMs), and apply them to train a neural network classifier independent of pre-selection methods. The halo images generated by DDPMs are of higher quality than those produced by WGANs. The diffusion-supported classifier with a multi-head attention block achieved the best average validation accuracy of 95.93% over 10 runs, using 36 clusters for training and 10 for testing, without further hyperparameter tuning. Using our classifier, we rediscovered 9/12 halos (75% detection rate) from the MeerKAT Galaxy Cluster Legacy Survey (MGCLS) Catalogue, and 5/8 halos (63% detection rate) from the Planck Sunyaev-Zeldovich Catalogue 2 (PSZ2) within the GaLactic and Extragalactic All-sky MWA (GLEAM) survey. In addition, we identify 11 potential new halos, minihalos, or candidates in the COSMOS field using XMM-chandra-detected clusters in GLEAM data. This work demonstrates the potential of ML for unbiased detection of diffuse emission and provides labeled datasets for further study.

Figures

Figures reproduced from arXiv: 2411.15559 by the authors.

Figure 1
Figure 1. DDPM for Image Generation: Starting from a highly noisy sample 𝑥𝑇, the model iteratively denoises the sample through intermediate states 𝑥𝑡 and 𝑥𝑡−1 using the reverse transition probability 𝑝(𝑥𝑡−1 | 𝑥𝑡 ), until it recon￾structs the original data 𝑥0. The dashed line indicates the forward process 𝑝(𝑥𝑡 | 𝑥𝑡−1 ), where noise is gradually added in the forward diffusion. the generator G, tries to create samples. The other… view at source ↗
Figure 2
Figure 2. Schematic representation ofWGAN used : The dots in the Generated Data box indicate that the generated image 𝐺(𝒛 | 𝜃 ) is a sample of the whole generated image distribution (𝑝𝐺 (𝒙| 𝜃 ) ) given by LWGAN-GP(𝐺, 𝐷; 𝜃) = Ex∼𝑝(x) [𝐷(x|𝜃)] −Ez∼𝑝z (z) [𝐷(𝐺(z|𝜃)|𝜃)]+ 𝜆Ex˜∼𝑝x˜ h (∥∇x˜ 𝐷(x˜|𝜃) ∥2 − 1) 2 i (6) In Eq (6), x˜ = 𝛼x+ (1−𝛼)𝐺(z|𝜃) with a random number 𝛼 ∼ 𝑈[0, 1] is an interpolated sample, and the factor 𝜆 determines … view at source ↗
Figure 4
Figure 4. An example of each type of ’halo’ (in the central field of view, in the frequency range 170-231 MHz, with an angular extent of 29.86 arcmin × 29.86 arcmin) used in the dataset; From top left to bottom left in a clockwise direction - Abell 0141 (H), Abell 2693 (cH), Abell 0122 (cmH) and rxcj0137 (mH). The colour scales are all in units of Jy beam−1 . 72-103 MHz, 103-134 MHz, 139-170 MHz, and 170-231 MHz. This allows … view at source ↗
Figures from the paper (8 more)
Figure 5
Figure 5. Figure 5: Pre-processing procedure for halo images. From left to right, the first panel shows an MWA-observed radio image. The second panel exhibits the image histogram with the salt-pepper background noise apparent towards the extreme left. The third panel of [PITH_FULL_IMAGE:…
Figure 6
Figure 6. Figure 6: The top panel shows the halo sources’ Largest Angular Size (LAS, green dots) against redshift, with the red dashed line marking the fixed angular extent of the 64x64 cut-out. The bottom panel compares the linear extent of the cut-outs (blue) to the sources’ Largest Lin…
Figure 7
Figure 7. Figure 7: Diffusion Model Evaluation: Pixel value distributions for diffusion-generated (blue) and real multi-frequency (orange) halo images are shown, with per-bin relative error in the bottom left panel. On the right, augmented unconditional diffusion-based images are displaye…
Figure 8
Figure 8. Figure 8: Validation losses during the training of (i) the base classifier {I, A, D} ( [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Receiver Operating Characteristic (ROC) Curves (true positive rate against the false positive rate) for Classifier A ( [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Confusion Matrix normalized over rows for Classifier 𝐴 ( [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 11
Figure 11. Figure 11: Pixel value distributions for generated (blue) and real images (orange), with per-bin relative error in the bottom panels for WGANs (left) and DDPMs (right) trained on Single-Frequency Images [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]
Figure 12
Figure 12. Figure 12: Set of augmented Real and GAN Generated images side by side (left); Set of augmented unconditional diffusion-based generated images (right) MNRAS 000, 1–14 (0000) [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Simulating realistic radio continuum survey maps with diffusion models

    astro-ph.IM 2025-06 conditional novelty 6.0 of 10

    A diffusion-model pipeline generates realistic simulated LOFAR survey maps with controllable galaxy sizes, closely matching real flux and size distributions.

  2. A Guided Unconditional Diffusion Model to Synthesize and Inpaint Radio Galaxies from FIRST, MGCLS and Radio Zoo

    astro-ph.GA 2026-01 conditional novelty 5.0 of 10

    A masked-guided diffusion model generates and inpaints radio galaxy images from a combined FIRST, MGCLS, and Radio Galaxy Zoo dataset.

Reference graph

Works this paper leans on

94 extracted references · 14 canonical work pages · cited by 2 Pith papers

  1. [1]

    arXiv:1603.04467

    Abadi M., et al., 2016, @doi [arXiv e-prints] 10.48550/arXiv.1603.04467 , https://ui.adsabs.harvard.edu/abs/2016arXiv160304467A p. arXiv:1603.04467

    arXiv doi:10.48550/arxiv.1603.04467 2016

  2. [2]

    Ackermann M., et al., 2014, @doi [The Astrophysical Journal] 10.1088/0004-637x/787/1/18 , 787, 18

    doi:10.1088/0004-637x/787/1/18 2014

  3. [3]

    Adam R., et al., 2017, @doi [ ] 10.1051/0004-6361/201629810 , https://ui.adsabs.harvard.edu/abs/2017A&A...606A..64A 606, A64

    doi:10.1051/0004-6361/201629810 2017

  4. [4]

    R., Vaccari M., 2018, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/sty2038 , 480, 2085

    Alhassan W., Taylor A. R., Vaccari M., 2018, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/sty2038 , 480, 2085

    doi:10.1093/mnras/sty2038 2018

  5. [5]

    Ananthakrishnan S., 1995, Journal of Astrophysics and Astronomy Supplement, https://ui.adsabs.harvard.edu/abs/1995JApAS..16..427A 16, 427

    1995

  6. [6]

    Andrade-Santos F., et al., 2017, @doi [The Astrophysical Journal] 10.3847/1538-4357/aa7461 , 843, 76

    doi:10.3847/1538-4357/aa7461 2017

  7. [7]

    K., Thorat K., 2017, @doi [ ] 10.3847/1538-4365/aa7333 , https://ui.adsabs.harvard.edu/abs/2017ApJS..230...20A 230, 20

    Aniyan A. K., Thorat K., 2017, @doi [ ] 10.3847/1538-4365/aa7333 , https://ui.adsabs.harvard.edu/abs/2017ApJS..230...20A 230, 20

    doi:10.3847/1538-4365/aa7333 2017

  8. [8]

    pp 214--223

    Arjovsky M., Chintala S., Bottou L., 2017, in Proceedings of the 34th International Conference on Machine Learning (ICML 2017). pp 214--223

    2017

Show all 94 references
  1. [9]

    Arora S., Zhang Y., 2017, Do GANs actually learn the distribution? An empirical study ( @eprint arXiv 1706.08224 )

    2017 arXiv

  2. [10]

    Bairouk A., Chaumont M., Fouchez D., Paquet J., Comby F., Bautista J., 2023, @doi [ ] 10.1051/0004-6361/202244657 , https://ui.adsabs.harvard.edu/abs/2023A&A...673A.141B 673, A141

    2023 doi

  3. [11]

    Balakrishnan V., Champion D., Barr E., Kramer M., Sengar R., Bailes M., 2021, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stab1308 , 505, 1180

    2021 doi

  4. [12]

    J., Scaife A

    Bastien D. J., Scaife A. M. M., Tang H., Bowles M., Porter F., 2021, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stab588 , 503, 3351

    2021 doi

  5. [13]

    et al., 2024, @doi [A&A] 10.1051/0004-6361/202348045 , 686, A82

    Biava, N. et al., 2024, @doi [A&A] 10.1051/0004-6361/202348045 , 686, A82

    2024 doi

  6. [14]

    J., Arbel M., Gretton A., 2018, in International Conference on Learning Representations

    Bińkowski M., Sutherland D. J., Arbel M., Gretton A., 2018, in International Conference on Learning Representations

    2018

  7. [15]

    B., 2010, @doi [ ] 10.1051/0004-6361/200913696 , https://ui.adsabs.harvard.edu/abs/2010A&A...513A..30B 513, A30

    Bonafede A., Feretti L., Murgia M., Govoni F., Giovannini G., Dallacasa D., Dolag K., Taylor G. B., 2010, @doi [ ] 10.1051/0004-6361/200913696 , https://ui.adsabs.harvard.edu/abs/2010A&A...513A..30B 513, A30

    2010 doi

  8. [16]

    Boone F., 2013, @doi [Experimental Astronomy] 10.1007/s10686-012-9322-1 , https://ui.adsabs.harvard.edu/abs/2013ExA....36...77B 36, 77

    2013 doi

  9. [17]

    Botteon A., et al., 2022, @doi [ ] 10.1051/0004-6361/202143020 , https://ui.adsabs.harvard.edu/abs/2022A&A...660A..78B 660, A78

    2022 doi

  10. [18]

    Bowles M., Bromley M., Allen M., Scaife A., 2021, E(2) Equivariant Self-Attention for Radio Astronomy ( @eprint arXiv 2111.04742 ), https://arxiv.org/abs/2111.04742

    2021 arXiv

  11. [19]

    S., 1995, in American Astronomical Society Meeting Abstracts

    Briggs D. S., 1995, in American Astronomical Society Meeting Abstracts. p. 112.02

    1995

  12. [20]

    W., 2014, International Journal of Modern Physics D, 23, 1430007

    Brunetti G., Jones T. W., 2014, International Journal of Modern Physics D, 23, 1430007

    2014

  13. [21]

    Brunetti G., Lazarian A., 2016, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stw496 , 458, 2584

    2016 doi

  14. [22]

    Brunetti G., Zimmer S., Zandanel F., 2017, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stx2092 , 472, 1506

    2017 doi

  15. [23]

    M., Scaife A

    Cantwell T. M., Scaife A. M. M., Oozeer N., Wen Z. L., Han J. L., 2016, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stw419 , 458, 1803

    2016 doi

  16. [24]

    Cassano R., Ettori S., Giacintucci S., Brunetti G., Markevitch M., Venturi T., Gitti M., 2010, @doi [ ] 10.1088/2041-8205/721/2/L82 , https://ui.adsabs.harvard.edu/abs/2010ApJ...721L..82C 721, L82

    2010 doi

  17. [25]

    E., Kronberg P

    Clarke T. E., Kronberg P. P., B \"o hringer H., 2001, @doi [ ] 10.1086/318896 , https://ui.adsabs.harvard.edu/abs/2001ApJ...547L.111C 547, L111

    2001 doi

  18. [26]

    Cuciti V., Cassano R., Brunetti G., Dallacasa D., Kale R., Ettori S., Venturi T., 2015, @doi [ ] 10.1051/0004-6361/201526420 , https://ui.adsabs.harvard.edu/abs/2015A&A...580A..97C 580, A97

    2015 doi

  19. [27]

    arXiv:2010.11929

    Dosovitskiy A., et al., 2020, @doi [arXiv e-prints] 10.48550/arXiv.2010.11929 , https://ui.adsabs.harvard.edu/abs/2020arXiv201011929D p. arXiv:2010.11929

    2020 doi arXiv

  20. [28]

    W., Johnston-Hollitt M., Bartalucci I., 2021b, @doi [Publications of the Astronomical Society of Australia] 10.1017/pasa.2021.45 , 38

    Duchesne S. W., Johnston-Hollitt M., Bartalucci I., 2021b, @doi [Publications of the Astronomical Society of Australia] 10.1017/pasa.2021.45 , 38

    2021 doi

  21. [29]

    W., Johnston-Hollitt M., Wilber A

    Duchesne S. W., Johnston-Hollitt M., Wilber A. G., 2021a, @doi [Publications of the Astronomical Society of Australia] 10.1017/pasa.2021.24 , 38

    2021 doi

  22. [30]

    W., Johnston-Hollitt M., Offringa A

    Duchesne S. W., Johnston-Hollitt M., Offringa A. R., Pratt G. W., Zheng Q., Dehghan S., 2021c, @doi [ ] 10.1017/pasa.2021.7 , https://ui.adsabs.harvard.edu/abs/2021PASA...38...10D 38, e010

    2021 doi

  23. [31]

    Feretti L., Giovannini G., Govoni F., Murgia M., 2012, Astronomy & Astrophysics Review, 20, 54

    2012

  24. [32]

    Forman W., Jones C., 1982, @doi [ ] 10.1146/annurev.aa.20.090182.002555 , https://ui.adsabs.harvard.edu/abs/1982ARA&A..20..547F 20, 547

    1982

  25. [33]

    T., et al., 2017, @doi [ ] 10.1093/mnras/stx155 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.467..936G 467, 936

    George L. T., et al., 2017, @doi [ ] 10.1093/mnras/stx155 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.467..936G 467, 936

    2017 doi

  26. [34]

    E., Brunetti G., 2017, @doi [ ] 10.3847/1538-4357/aa7069 , https://ui.adsabs.harvard.edu/abs/2017ApJ...841...71G 841, 71

    Giacintucci S., Markevitch M., Cassano R., Venturi T., Clarke T. E., Brunetti G., 2017, @doi [ ] 10.3847/1538-4357/aa7069 , https://ui.adsabs.harvard.edu/abs/2017ApJ...841...71G 841, 71

    2017 doi

  27. [35]

    Giovannini G., Tordi M., Feretti L., 1999, @doi [ ] 10.1016/S1384-1076(99)00018-4 , https://ui.adsabs.harvard.edu/abs/1999NewA....4..141G 4, 141

    1999 doi

  28. [36]

    Goodfellow I., Pouget-Abadie J., Mirza M., Xu B., Warde-Farley D., Ozair S., Courville A., Bengio Y., 2014, Advances in neural information processing systems, 27

    2014

  29. [37]

    C., 2017, Advances in neural information processing systems, 30

    Gulrajani I., Ahmed F., Arjovsky M., Dumoulin V., Courville A. C., 2017, Advances in neural information processing systems, 30

    2017

  30. [38]

    Hackstein S., Kinakh V., Bailer C., Melchior M., 2023, @doi [Astronomy and Computing] https://doi.org/10.1016/j.ascom.2022.100685 , 42, 100685

    2023

  31. [39]

    pp 6626--6637

    Heusel M., Ramsauer H., Unterthiner T., Nessler B., Hochreiter S., 2017, in Advances in Neural Information Processing Systems. pp 6626--6637

    2017

  32. [40]

    arXiv:2006.11239

    Ho J., Jain A., Abbeel P., 2020, @doi [arXiv e-prints] 10.48550/arXiv.2006.11239 , https://ui.adsabs.harvard.edu/abs/2020arXiv200611239H p. arXiv:2006.11239

    2020 doi arXiv

  33. [41]

    W., et al., 2021, @doi [ ] 10.1017/pasa.2021.1 , https://ui.adsabs.harvard.edu/abs/2021PASA...38....9H 38, e009

    Hotan A. W., et al., 2021, @doi [ ] 10.1017/pasa.2021.1 , https://ui.adsabs.harvard.edu/abs/2021PASA...38....9H 38, e009

    2021 doi

  34. [42]

    T., Baumann S

    Hu V. T., Baumann S. A., Gui M., Grebenkova O., Ma P., Fischer J., Ommer B., 2024, @doi [arXiv e-prints] 10.48550/arXiv.2403.13802 , https://ui.adsabs.harvard.edu/abs/2024arXiv240313802H p. arXiv:2403.13802

    2024 doi arXiv

  35. [43]

    Hurley-Walker N., et al., 2017, @doi [ ] 10.1093/mnras/stw2337 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.464.1146H 464, 1146

    2017 doi

  36. [44]

    Springer, New York

    James G., Witten D., Hastie T., Tibshirani R., 2013, An Introduction to Statistical Learning: with Applications in R. Springer, New York

    2013

  37. [45]

    Jonas J., MeerKAT Team 2016, in MeerKAT Science: On the Pathway to the SKA. p. 1, @doi 10.22323/1.277.0001

    2016 doi

  38. [46]

    P., Ba J., 2017, Adam: A Method for Stochastic Optimization ( @eprint arXiv 1412.6980 )

    Kingma D. P., Ba J., 2017, Adam: A Method for Stochastic Optimization ( @eprint arXiv 1412.6980 )

    2017 arXiv

  39. [47]

    Knowles K., et al., 2021, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stab939 , 504, 1749

    2021 doi

  40. [48]

    Knowles K., et al., 2022, @doi [ ] 10.1051/0004-6361/202141488 , https://ui.adsabs.harvard.edu/abs/2022A&A...657A..56K 657, A56

    2022 doi

  41. [49]

    V., Borgani S., 2012, @doi [ ] 10.1146/annurev-astro-081811-125502 , https://ui.adsabs.harvard.edu/abs/2012ARA&A..50..353K 50, 353

    Kravtsov A. V., Borgani S., 2012, @doi [ ] 10.1146/annurev-astro-081811-125502 , https://ui.adsabs.harvard.edu/abs/2012ARA&A..50..353K 50, 353

    2012 doi

  42. [50]

    arXiv:2206.15131

    Kummer J., et al., 2022, @doi [arXiv e-prints] 10.48550/arXiv.2206.15131 , https://ui.adsabs.harvard.edu/abs/2022arXiv220615131K p. arXiv:2206.15131

    2022 doi arXiv

  43. [51]

    Lecun Y., Bottou L., Bengio Y., Haffner P., 1998, @doi [Proceedings of the IEEE] 10.1109/5.726791 , 86, 2278

    1998 doi

  44. [52]

    Lukic V., de Gasperin F., Br \"u ggen M., 2019, @doi [Galaxies] 10.3390/galaxies8010003 , https://ui.adsabs.harvard.edu/abs/2019Galax...8....3L 8, 3

    2019 doi

  45. [53]

    arXiv:1806.00398

    Ma Z., Zhu J., Li W., Xu H., 2018, @doi [arXiv e-prints] 10.48550/arXiv.1806.00398 , https://ui.adsabs.harvard.edu/abs/2018arXiv180600398M p. arXiv:1806.00398

    2018 doi arXiv

  46. [54]

    Maslej-Kre s n \'a kov \'a V., El Bouchefry K., Butka P., 2021, @doi [ ] 10.1093/mnras/stab1400 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.505.1464M 505, 1464

    2021 doi

  47. [55]

    N., 2017, @doi [ ] 10.1093/mnras/stx007 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.466.4346M 466, 4346

    Miraghaei H., Best P. N., 2017, @doi [ ] 10.1093/mnras/stx007 , https://ui.adsabs.harvard.edu/abs/2017MNRAS.466.4346M 466, 4346

    2017 doi

  48. [56]

    Mirza M., Osindero S., 2014, Conditional Generative Adversarial Nets ( @eprint arXiv 1411.1784 )

    2014 arXiv

  49. [57]

    J., Culhane J

    Mitchell R. J., Culhane J. L., Davison P. J. N., Ives J. C., 1976, @doi [ ] 10.1093/mnras/175.1.29P , https://ui.adsabs.harvard.edu/abs/1976MNRAS.175P..29M 175, 29P

    1976 doi

  50. [58]

    F., 1950, @doi [Proceedings of the National Academy of Science] 10.1073/pnas.36.1.48 , https://ui.adsabs.harvard.edu/abs/1950PNAS...36...48N 36, 48

    Nash J. F., 1950, @doi [Proceedings of the National Academy of Science] 10.1073/pnas.36.1.48 , https://ui.adsabs.harvard.edu/abs/1950PNAS...36...48N 36, 48

    1950 doi

  51. [59]

    E., 2021, @doi [ ] 10.1093/mnras/stab271 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.502.3417N 502, 3417

    Ntwaetsile K., Geach J. E., 2021, @doi [ ] 10.1093/mnras/stab271 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.502.3417N 502, 3417

    2021 doi

  52. [60]

    Paul S., et al., 2023, @doi [Journal of Astrophysics and Astronomy] 10.1007/s12036-023-09914-2 , 44

    2023 doi

  53. [61]

    B., Babul A., McCarthy I

    Poole G. B., Babul A., McCarthy I. G., Fardal M. A., Bildfell C. J., Quinn T., Mahdavi A., 2007, @doi [Monthly Notices of the Royal Astronomical Society] 10.1111/j.1365-2966.2007.12107.x , 380, 437

    2007

  54. [62]

    Radford A., Metz L., Chintala S., 2016, Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks ( @eprint arXiv 1511.06434 )

    2016 arXiv

  55. [63]

    J., Intema H

    Raja R., Rahaman M., Datta A., van Weeren R. J., Intema H. T., Paul S., 2020, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/staa3432 , 500, 2236–2249

    2020 doi

  56. [64]

    V., 2017, @doi [arXiv e-prints] 10.48550/arXiv.1710.05941 , https://ui.adsabs.harvard.edu/abs/2017arXiv171005941R p

    Ramachandran P., Zoph B., Le Q. V., 2017, @doi [arXiv e-prints] 10.48550/arXiv.1710.05941 , https://ui.adsabs.harvard.edu/abs/2017arXiv171005941R p. arXiv:1710.05941

    2017 doi arXiv

  57. [65]

    W., Sarazin C

    Randall S. W., Sarazin C. L., Ricker P. M., 2002, @doi [The Astrophysical Journal] 10.1086/342239 , 577, 579

    2002 doi

  58. [66]

    arXiv:2112.10752

    Rombach R., Blattmann A., Lorenz D., Esser P., Ommer B., 2021, @doi [arXiv e-prints] 10.48550/arXiv.2112.10752 , https://ui.adsabs.harvard.edu/abs/2021arXiv211210752R p. arXiv:2112.10752

    2021 doi arXiv

  59. [67]

    arXiv:2208.12242

    Ruiz N., Li Y., Jampani V., Pritch Y., Rubinstein M., Aberman K., 2022, @doi [arXiv e-prints] 10.48550/arXiv.2208.12242 , https://ui.adsabs.harvard.edu/abs/2022arXiv220812242R p. arXiv:2208.12242

    2022 doi arXiv

  60. [68]

    Rustige L., et al., 2023, @doi [RAS Techniques and Instruments] 10.1093/rasti/rzad016 , 2, 264

    2023 doi

  61. [69]

    Sadeghi M., Javaherian M., Miraghaei H., 2021, @doi [The Astronomical Journal] 10.3847/1538-3881/abd314 , 161, 94

    2021 doi

  62. [70]

    T., Bansal M., Wadadekar Y., 2022, @doi [ ] 10.1093/mnras/stab3144 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.509.2269S 509, 2269

    Samudre A., George L. T., Bansal M., Wadadekar Y., 2022, @doi [ ] 10.1093/mnras/stab3144 , https://ui.adsabs.harvard.edu/abs/2022MNRAS.509.2269S 509, 2269

    2022 doi

  63. [71]

    Sanderson A., Ponman T., Finoguenov A., Lloyd-Davies E., Markevitch M., 2003, @doi [Monthly Notices of the Royal Astronomical Society] 10.1046/j.1365-8711.2003.06401.x , 340, 989

    2003

  64. [72]

    Scaife A. M. M., Porter F., 2021, @doi [ ] 10.1093/mnras/stab530 , https://ui.adsabs.harvard.edu/abs/2021MNRAS.503.2369S 503, 2369

    2021 doi

  65. [73]

    K., 2017, @doi [Monthly Notices of the Royal Astronomical Society: Letters] 10.1093/mnrasl/slx008 , 467, L110

    Schawinski K., Zhang C., Zhang H., Fowler L., Santhanam G. K., 2017, @doi [Monthly Notices of the Royal Astronomical Society: Letters] 10.1093/mnrasl/slx008 , 467, L110

    2017 doi

  66. [74]

    Schlickeiser R., Sievers A., Thiemann H., 1987, Astronomy and Astrophysics (ISSN 0004-6361), vol. 182, no. 1, Aug. 1987, p. 21-35., 182, 21

    1987

  67. [75]

    J., Smith B

    Serlemitsos P. J., Smith B. W., Boldt E. A., Holt S. S., Swank J. H., 1977, @doi [ ] 10.1086/182342 , https://ui.adsabs.harvard.edu/abs/1977ApJ...211L..63S 211, L63

    1977 doi

  68. [76]

    Smithsonian Astrophysical Observatory 2000, SAOImage DS9: A Utility for Displaying Astronomical Images in the X11 Window Environment, Astrophysics Source Code Library, record ascl:0003.002, https://ui.adsabs.harvard.edu/abs/2000ascl.soft03002S

    2000

  69. [77]

    37, Proceedings of the 32nd International Conference on Machine Learning

    Sohl-Dickstein J., Weiss E., Maheswaranathan N., Ganguli S., 2015, in Bach F., Blei D., eds, Proceedings of Machine Learning Research Vol. 37, Proceedings of the 32nd International Conference on Machine Learning. PMLR, Lille, France, pp 2256--2265, https://proceedings.mlr.pres...

    2015

  70. [78]

    arXiv:2307.02392

    Sortino R., et al., 2023, @doi [arXiv e-prints] 10.48550/arXiv.2307.02392 , https://ui.adsabs.harvard.edu/abs/2023arXiv230702392S p. arXiv:2307.02392

    2023 doi arXiv

  71. [79]

    A., Zeldovich Y

    Sunyaev R. A., Zeldovich Y. B., 1972, Comments on Astrophysics and Space Physics, https://ui.adsabs.harvard.edu/abs/1972CoASP...4..173S 4, 173

    1972

  72. [80]

    pp 2818--2826

    Szegedy C., Vanhoucke V., Ioffe S., Shlens J., Wojna Z., 2016, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp 2818--2826

    2016

  73. [81]

    Tang H., Scaife A. M. M., Leahy J. P., 2019, @doi [ ] 10.1093/mnras/stz1883 , https://ui.adsabs.harvard.edu/abs/2019MNRAS.488.3358T 488, 3358

    2019 doi

  74. [82]

    J., et al., 2013, @doi [ ] 10.1017/pasa.2012.007 , https://ui.adsabs.harvard.edu/abs/2013PASA...30....7T 30, e007

    Tingay S. J., et al., 2013, @doi [ ] 10.1017/pasa.2012.007 , https://ui.adsabs.harvard.edu/abs/2013PASA...30....7T 30, e007

    2013 doi

  75. [83]

    Butterworth-Heinemann, London

    Van Rijsbergen C., 1979, Information Retrieval. Butterworth-Heinemann, London

    1979

  76. [84]

    N., Kaiser L., Polosukhin I., 2017, @doi [arXiv e-prints] 10.48550/arXiv.1706.03762 , https://ui.adsabs.harvard.edu/abs/2017arXiv170603762V p

    Vaswani A., Shazeer N., Parmar N., Uszkoreit J., Jones L., Gomez A. N., Kaiser L., Polosukhin I., 2017, @doi [arXiv e-prints] 10.48550/arXiv.1706.03762 , https://ui.adsabs.harvard.edu/abs/2017arXiv170603762V p. arXiv:1706.03762

    2017 doi arXiv

  77. [85]

    S., Van Speybroeck L., 2006, @doi [ ] 10.1086/500288 , https://ui.adsabs.harvard.edu/abs/2006ApJ...640..691V 640, 691

    Vikhlinin A., Kravtsov A., Forman W., Jones C., Markevitch M., Murray S. S., Van Speybroeck L., 2006, @doi [ ] 10.1086/500288 , https://ui.adsabs.harvard.edu/abs/2006ApJ...640..691V 640, 691

    2006 doi

  78. [86]

    B., et al., 2015, @doi [ ] 10.1017/pasa.2015.26 , https://ui.adsabs.harvard.edu/abs/2015PASA...32...25W 32, e025

    Wayth R. B., et al., 2015, @doi [ ] 10.1017/pasa.2015.26 , https://ui.adsabs.harvard.edu/abs/2015PASA...32...25W 32, e025

    2015 doi

  79. [87]

    R., Sarazin C

    Wik D. R., Sarazin C. L., Ricker P. M., Randall S. W., 2008, @doi [The Astrophysical Journal] 10.1086/587790 , 680, 17

    2008 doi

  80. [88]

    G., Johnston-Hollitt M., Duchesne S

    Wilber A. G., Johnston-Hollitt M., Duchesne S. W., Tasse C., Akamatsu H., Intema H., Hodgson T., 2020, @doi [Publications of the Astronomical Society of Australia] 10.1017/pasa.2020.34 , 37, e040

    2020 doi

  81. [89]

    Wu D., Cao H., Lv N., Fan J., Tan X., Yang S., 2019, @doi [The Astrophysical Journal Letters] 10.3847/2041-8213/ab595e , 887, L10

    2019 doi

  82. [90]

    Zhang W., Li X., Chen Y., Wang X., 2023, arXiv preprint arXiv:2310.08442

    2023 arXiv

  83. [91]

    T., Frail D

    de Gasperin F., Intema H. T., Frail D. A., 2017, @doi [Monthly Notices of the Royal Astronomical Society] 10.1093/mnras/stx3125 , 474, 5008

    2017 doi

  84. [92]

    P., et al., 2013, @doi [ ] 10.1051/0004-6361/201220873 , https://ui.adsabs.harvard.edu/abs/2013A&A...556A...2V 556, A2

    van Haarlem M. P., et al., 2013, @doi [ ] 10.1051/0004-6361/201220873 , https://ui.adsabs.harvard.edu/abs/2013A&A...556A...2V 556, A2

    2013 doi

  85. [93]

    u ggen M., R \

    van Weeren R. J., Br \"u ggen M., R \"o ttgering H. J. A., Hoeft M., Nuza S. E., Intema H. T., 2011, @doi [ ] 10.1051/0004-6361/201117149 , https://ui.adsabs.harvard.edu/abs/2011A&A...533A..35V 533, A35

    2011 doi

  86. [94]

    van Weeren R. J., de Gasperin F., Akamatsu H., Br \"u ggen M., Feretti L., Kang H., Stroe A., Zandanel F., 2019, @doi [ ] 10.1007/s11214-019-0584-z , https://ui.adsabs.harvard.edu/abs/2019SSRv..215...16V 215, 16

    2019 doi

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.