pith. sign in

arxiv: 2507.14854 · v2 · submitted 2025-07-20 · 🌌 astro-ph.IM · astro-ph.HE

Frabjous: Deep Learning Fast Radio Burst Morphologies

Ajay Kumar , Ashish A. Mahabal , Shriharsh P. Tendulkar This is my paper

Pith reviewed 2026-05-19 04:35 UTC · model grok-4.3

classification 🌌 astro-ph.IM astro-ph.HE
keywords fast radio burstsdeep learningmorphology classificationCHIME/FRB catalogsimulation frameworkautomated classificationradio astronomy
0
0 comments X

The pith

Deep learning trained on simulated and real data classifies fast radio burst morphologies at 55 percent accuracy.

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

The paper introduces Frabjous, a deep learning framework designed to automatically classify the morphologies of fast radio bursts. Limited real-world examples for each type are supplemented by a simulation framework that generates realistic additional training instances. When applied to the first CHIME/FRB catalog, this mixed training approach yields an overall accuracy of around 55 percent across five classes, exceeding the 20 percent rate of random guessing. Such classification supports prioritizing follow-up observations for unusual bursts and enables statistical studies of FRB shapes. The authors acknowledge current limitations and outline paths for refinement.

Core claim

Frabjous applies deep learning to FRB morphology classification by combining simulated examples with real observations from the CHIME/FRB catalog, resulting in an overall classification accuracy of approximately 55 percent for five balanced classes during training.

What carries the argument

The Frabjous deep learning model trained on a hybrid dataset of simulated FRB signals and real catalog entries.

If this is right

  • Automated morphology classification can help prioritize limited multi-wavelength follow-up resources for the most interesting FRBs.
  • Statistical analysis of FRB morphologies becomes feasible with growing detection rates.
  • Performance can be improved by augmenting training datasets and refining simulation strategies.

Where Pith is reading between the lines

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

  • Similar simulation-augmented training could benefit classification tasks in other areas with sparse labeled astronomical data.
  • Extending the model to handle more classes or continuous morphology parameters might reveal new patterns in FRB populations.
  • Integration with real-time detection pipelines could enable immediate anomaly flagging as new surveys increase event rates.

Load-bearing premise

The simulations used to generate additional training examples accurately capture the diversity and characteristics of real FRB morphologies observed in the CHIME/FRB catalog.

What would settle it

Evaluating the trained model on a held-out set of newly observed real FRBs with independently determined morphologies and checking whether the accuracy remains above 50 percent.

Figures

Figures reproduced from arXiv: 2507.14854 by Ajay Kumar, Ashish A. Mahabal, Shriharsh P. Tendulkar.

Figure 1
Figure 1. Figure 1: In each row from left to right example of type I, II, III, IV, V and VI burst morphology simulated from our framework. We have randomly chosen three bursts from simulations for each type to demonstrate the different types. 2.3. Caveats of the simulation framework While we have addressed most of the features cur￾rently observed in published FRBs, there are still some limitations to the simulated training da… view at source ↗
Figure 3
Figure 3. Figure 3: A schematic diagram illustrating the workflow where input dynamic spectra is processed through multiple binary classifiers (with the first class labeled as negative and the second as positive). The outputs from these binary classi￾fiers are then combined to infer the final classification, follow￾ing the multi-classification framework detailed in Section 4.4. For clarity, only a subset of binary classifiers… view at source ↗
Figure 4
Figure 4. Figure 4: A schematic diagram of a typical binary classifier network. A dynamic spectrum (256 × 256) is the input for the classifier. The first few layers are convolutional, the next layers are fully connected layers, and the final layer is one that gives the confidence of the input belonging to the classes under consideration. Figure made using NN-SVG (LeNail 2019). 0 250 500 750 1000 1250 1500 1750 2000 epoch 0.4 … view at source ↗
Figure 5
Figure 5. Figure 5: An instance of training a specific network with over 2000 epochs for binary classification of type IV vs type V. Left Panel: Accuracy vs. epoch for the training (blue) and validation set (orange). Right Panel: Loss (binary cross-entropy loss function) as a function of epoch. In [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: We present pairwise binary classification confusion matrices. Each confusion matrix represents the inference on the test data (includes samples for all SNR values) from simulated dataset for an optimised model obtained by hyperparameter tuning for each of the binary classification. Light gray boxes represent the correct classifications and dark gray represent the incorrect classifications. Top number in ea… view at source ↗
Figure 7
Figure 7. Figure 7: False positive rate (FPR) and false negative rate (FNR) as a function of confidence output of the classi￾fier. The intersection of the FPR and FNR curves signifies the value of confidence at which false positives equal false negatives. This specific example is for a type I vs II classi￾fier. and the second dense layer, learning rate, batch size, and dropout. Range of values taken for these hyperpa￾rameters… view at source ↗
Figure 8
Figure 8. Figure 8: Example classification matrix for a single burst showing the output confidence from all the optimised binary models. Each element indicates the output score from a bi￾nary classifier corresponding to the archetype denoted by the row and column. Each element’s upper value corresponds to the confidence i.e. the output confidence of the classifier while the lower value represents the optimal thresholds de￾ter… view at source ↗
Figure 9
Figure 9. Figure 9: Each violin plot in this figure displays distribution of the augmented confidences sum for one type as described in [PITH_FULL_IMAGE:figures/full_fig_p013_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: a) This figure shows the Intensity as a function of frequency channels for one particular time sample corresponding to FRB emission seen in dynamic spectra of one of the CHIME/FRB catalog. The masked channels are indicated by orange hashes. The pink hashes indicate the neighbouring region used to fill the masked channels. b) Dynamic spectra of one of the CHIME/FRB burst c) Dynamic spectra after using line… view at source ↗
Figure 12
Figure 12. Figure 12: Confusion matrix after classifying the CHIME/FRB first catalog using the multi-class framework described in Section 4.4. In each square, the upper num￾ber corresponds to the bursts classified in that combination of true (row) and predicted (column) archetypes. The lower number is the same as a fraction of the total number of bursts of that type in the CHIME/FRB catalog — 163, 270, 62, 20, 20 samples for t… view at source ↗
Figure 11
Figure 11. Figure 11: a) Waterfall plot of one of the type II CHIME/FRB catalog bursts. b) Model waterfall plot for the same burst shown in panel a). c) This matrix is as de￾scribed in [PITH_FULL_IMAGE:figures/full_fig_p015_11.png] view at source ↗
Figure 13
Figure 13. Figure 13: Confusion matrix for multi class classification using a single multi-class classifier. The details of the plot are the same as in [PITH_FULL_IMAGE:figures/full_fig_p015_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Left Panel: Confusion matrix with half of CHIME/FRB test catalog for the type II, III, IV, and V after we get optimised model as described in the text. In each element, the value above is the number corresponding to predicted archetype and actual archetype. The values below is the percentage of the particular type. Middle Panel: Similar to the left panel, the confusion matrix for type I, III, IV, V. Right… view at source ↗
Figure 15
Figure 15. Figure 15: Left Panel: Same as [PITH_FULL_IMAGE:figures/full_fig_p019_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: SHAP values for three correctly classified burst examples from the CHIME/FRB catalogue. In each row, the left column shows the preprocessed input waterfall and the next five columns show SHAP value outputs of each classifier, ordered in decreasing order of classification confidence. In each SHAP plot, the pixels in red (blue) contribute positively (negatively) to the prediction of the model. From top to b… view at source ↗
Figure 17
Figure 17. Figure 17: Same plots as [PITH_FULL_IMAGE:figures/full_fig_p020_17.png] view at source ↗
read the original abstract

The increasing field of view of radio telescopes and improved data processing capabilities have led to a surge in the detection of Fast Radio Bursts (FRBs). The discovery rate of FRBs is already a few per day and is expected to increase rapidly with new surveys coming online. The growing number of events necessitates prioritized follow-up due to limited multi-wavelength resources, requiring rapid and automated classification. In this study, we introduce Frabjous, a deep learning framework for an automated morphology classifier with an aim towards enabling the prompt follow-up of anomalous and intriguing FRBs, and a comprehensive statistical analysis of FRB morphologies. Deep learning models require a large training set of each FRB archetype, however, publicly available data lacks sufficient samples for most FRB types. In this paper, we build a simulation framework for generating realistic examples of FRBs and train a network based on a combination of simulated and real data, starting with the CHIME/FRB catalog. Applying our framework to the first CHIME/FRB catalog, we achieve an overall classification accuracy of approximately 55%, well over a random multiclass classification rate of 20 % with five balanced classes during training. While this falls short of desirable performance, we critically discuss the limitations of our approach and propose potential avenues for improvement. Future work should explore strategies to augment training datasets and broaden the scope of FRB morphological studies, aiming for more accurate and reliable classification results.

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

1 major / 1 minor

Summary. The manuscript introduces Frabjous, a deep learning framework for automated classification of Fast Radio Burst (FRB) morphologies. To overcome limited labeled real data, the authors develop a simulation framework to generate additional training examples and combine these with real events from the first CHIME/FRB catalog. They train a network on the augmented dataset for five morphology classes and report an overall classification accuracy of approximately 55%, exceeding the 20% random baseline. The authors explicitly note that performance falls short of desirable levels, discuss limitations, and outline potential improvements for dataset augmentation and broader morphological studies.

Significance. If the simulations are shown to reproduce the statistical properties of real FRB morphologies, the framework could provide a practical route to automated classification tools that help prioritize limited multi-wavelength follow-up resources as detection rates rise. The strategy of augmenting scarce real data with targeted simulations directly addresses a common bottleneck in radio transient studies and supplies a concrete baseline (55% vs. 20%) against which future refinements can be measured.

major comments (1)
  1. [Simulation framework and training procedure] The central claim of ~55% accuracy on the CHIME/FRB catalog (abstract) requires that simulated training examples capture the morphological statistics (width, scattering, spectral structure, noise) of real events sufficiently well for the network to learn transferable features. No quantitative validation metrics—such as parameter-distribution overlap, expert visual scoring, or domain-adversarial distances—are reported for the five classes. Without these checks, the improvement over the random baseline could reflect simulation-specific artifacts rather than genuine generalization to the catalog.
minor comments (1)
  1. [Abstract] The abstract states that five balanced classes were used during training but does not name the morphological archetypes; adding the class labels would aid readers who are not already familiar with the FRB morphology taxonomy.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive and insightful comments on our manuscript. We have carefully considered the major comment and provide a detailed response below. We believe the suggested revisions will improve the rigor of our presentation of the simulation framework.

read point-by-point responses

  1. Referee: The central claim of ~55% accuracy on the CHIME/FRB catalog (abstract) requires that simulated training examples capture the morphological statistics (width, scattering, spectral structure, noise) of real events sufficiently well for the network to learn transferable features. No quantitative validation metrics—such as parameter-distribution overlap, expert visual scoring, or domain-adversarial distances—are reported for the five classes. Without these checks, the improvement over the random baseline could reflect simulation-specific artifacts rather than genuine generalization to the catalog.

    Authors: We agree that quantitative validation of the simulated morphologies against real events is essential to substantiate the transferability of learned features and to rule out simulation-specific artifacts. The current manuscript discusses limitations of the approach and notes that the 55% accuracy falls short of desirable performance, but does not include the specific metrics suggested. In the revised manuscript we will add direct comparisons of key parameter distributions (burst width, scattering timescale, spectral index, and noise properties) between simulated and real FRBs for each of the five morphology classes, including summary statistics and overlap measures such as Kolmogorov-Smirnov tests. We will also include a small-scale expert visual scoring exercise on a representative subset of simulated events. While full implementation of domain-adversarial distance metrics would require substantial additional computational work, we will expand the limitations and future-work sections to discuss this technique as a promising direction for further validation. These additions will be incorporated in the next version of the paper. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical ML accuracy on external catalog data

full rationale

The paper describes a standard supervised deep learning pipeline: a simulation framework generates additional training examples of FRB morphologies, which are combined with real CHIME/FRB catalog events to train a classifier. The reported ~55% accuracy is an empirical evaluation metric on the catalog, not the output of any equation or parameter fit that reduces to the training inputs by construction. No self-definitional steps, fitted inputs renamed as predictions, load-bearing self-citations, or ansatzes appear in the derivation chain. The result depends on data and model training rather than tautological re-expression of its own assumptions.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The central claim rests on the premise that simulated FRB signals can be generated with sufficient fidelity to supplement scarce real examples and produce a generalizable classifier; this is a domain assumption rather than a derived result.

free parameters (2)
  • Number of morphological classes
    Set to five balanced classes for training and evaluation as stated in the abstract.
  • Simulation hyperparameters
    Parameters controlling the generation of realistic FRB examples are required to produce the augmented training set but are not enumerated in the abstract.
axioms (1)
  • domain assumption Simulated FRB signals can be made sufficiently realistic to augment limited real data for training a morphology classifier.
    The abstract explicitly states that publicly available data lacks sufficient samples and therefore relies on a simulation framework to generate additional examples.

pith-pipeline@v0.9.0 · 5797 in / 1460 out tokens · 43120 ms · 2026-05-19T04:35:56.641763+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

72 extracted references · 72 canonical work pages · 3 internal anchors

  1. [1]

    2015, TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems

    Abadi, M., Agarwal, A., Barham, P., et al. 2015, TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems. https://www.tensorflow.org/

    work page 2015

  2. [2]

    2016, in 12th {USENIX} Symposium on Operating Systems Design and Implementation ({OSDI} 16), 265–283

    Abadi, M., Barham, P., Chen, J., et al. 2016, in 12th {USENIX} Symposium on Operating Systems Design and Implementation ({OSDI} 16), 265–283

    work page 2016

  3. [3]

    C., Buffaz, E., Vieira, N., et al

    Abbott, T. C., Buffaz, E., Vieira, N., et al. 2022, ApJ, 927, 232, doi: 10.3847/1538-4357/ac5019

    work page doi:10.3847/1538-4357/ac5019 2022

  4. [4]

    2020, FETCH: Fast Extragalactic Transient Candidate Hunter, Astrophysics Source Code Library, record ascl:2005.014

    Agarwal, D., & Aggarwal, K. 2020, FETCH: Fast Extragalactic Transient Candidate Hunter, Astrophysics Source Code Library, record ascl:2005.014. http://ascl.net/2005.014

    work page 2020

  5. [5]

    R., Garver-Daniels N., 2020, @doi [ ] 10.1093/mnras/staa1856 , https://ui.adsabs.harvard.edu/abs/2020MNRAS.497.1661A 497, 1661

    Agarwal, D., Aggarwal, K., Burke-Spolaor, S., Lorimer, D. R., & Garver-Daniels, N. 2020, MNRAS, 497, 1661, doi: 10.1093/mnras/staa1856

    work page doi:10.1093/mnras/staa1856 2020

  6. [6]

    2020, ApJ, 888, 40, doi: 10.3847/1538-4357/ab5363

    Balasubramanian, A. 2020, ApJ, 888, 40, doi: 10.3847/1538-4357/ab5363

    work page doi:10.3847/1538-4357/ab5363 2020

  7. [7]

    BEiT: BERT Pre-Training of Image Transformers

    Bao, H., Dong, L., Piao, S., & Wei, F. 2021, arXiv e-prints, arXiv:2106.08254, doi: 10.48550/arXiv.2106.08254

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2106.08254 2021

  8. [8]

    G., Tendulkar, S

    Bassa, C. G., Tendulkar, S. P., Adams, E. A. K., et al. 2017, ApJL, 843, L8, doi: 10.3847/2041-8213/aa7a0c

    work page doi:10.3847/2041-8213/aa7a0c 2017

  9. [9]

    M., Prochaska, J

    Bhandari, S., Sadler, E. M., Prochaska, J. X., et al. 2020, ApJL, 895, L37, doi: 10.3847/2041-8213/ab672e

    work page doi:10.3847/2041-8213/ab672e 2020

  10. [10]

    D., Ravi V., Belov K

    Bochenek, C. D., Ravi, V., Belov, K. V., et al. 2020, Nature, 587, 59, doi: 10.1038/s41586-020-2872-x

    work page doi:10.1038/s41586-020-2872-x 2020

  11. [11]

    2022, Nature Astronomy, 6, 828, doi: 10.1038/s41550-022-01688-x

    Caleb, M., Heywood, I., Rajwade, K., et al. 2022, Nature Astronomy, 6, 828, doi: 10.1038/s41550-022-01688-x

    work page doi:10.1038/s41550-022-01688-x 2022

  12. [12]

    2023, arXiv e-prints, arXiv:2307.09502, doi: 10.48550/arXiv.2307.09502

    Cassanelli, T., Leung, C., Sanghavi, P., et al. 2023, arXiv e-prints, arXiv:2307.09502, doi: 10.48550/arXiv.2307.09502

    work page doi:10.48550/arxiv.2307.09502 2023

  13. [13]

    2017, Nature, 541, 58, 10.1038/nature20797

    Chatterjee, S., Law, C. J., Wharton, R. S., et al. 2017, Nature, 541, 58, doi: 10.1038/nature20797

    work page doi:10.1038/nature20797 2017

  14. [14]

    H., Hashimoto, T., Goto, T., et al

    Chen, B. H., Hashimoto, T., Goto, T., et al. 2022, MNRAS, 509, 1227, doi: 10.1093/mnras/stab2994 —. 2023, MNRAS, 521, 5738, doi: 10.1093/mnras/stad930

    work page doi:10.1093/mnras/stab2994 2022

  15. [15]

    Chen, S., Shu, T., Zhao, H., & Tang, Y. Y. 2023, Knowledge-Based Systems, 278, 110881, doi: https://doi.org/10.1016/j.knosys.2023.110881 CHIME/FRB Collaboration, Amiri, M., Andersen, B. C., et al. 2021, ApJS, 257, 59, doi: 10.3847/1538-4365/ac33ab CHIME/FRB Collaboration, Andersen, B. C., Bandura, K., et al. 2023, ApJ, 947, 83, doi: 10.3847/1538-4357/acc6...

    work page doi:10.1016/j.knosys.2023.110881 2023

  16. [16]

    2022, Nature, 607, 256, doi: 10.1038/s41586-022-04841-8

    Bhardwaj, M., et al. 2022, Nature, 607, 256, doi: 10.1038/s41586-022-04841-8

    work page doi:10.1038/s41586-022-04841-8 2022

  17. [17]

    M., et al

    Cho, H., Macquart, J.-P., Shannon, R. M., et al. 2020, ApJL, 891, L38, doi: 10.3847/2041-8213/ab7824

    work page doi:10.3847/2041-8213/ab7824 2020

  18. [18]

    2015, Keras, GitHub

    Chollet, F., et al. 2015, Keras, GitHub. https://github.com/fchollet/keras

    work page 2015

  19. [19]

    2018, AJ, 156, 256, doi: 10.3847/1538-3881/aae649

    Connor, L., & van Leeuwen, J. 2018, AJ, 156, 256, doi: 10.3847/1538-3881/aae649

    work page doi:10.3847/1538-3881/aae649 2018

  20. [20]

    K., Deller, A

    Day, C. K., Deller, A. T., Shannon, R. M., et al. 2020, Monthly Notices of the Royal Astronomical Society, 497, 3335, doi: 10.1093/mnras/staa2138 21

    work page doi:10.1093/mnras/staa2138 2020

  21. [21]

    T., Michilli, D., Mckinven, R., et al

    Faber, J. T., Michilli, D., Mckinven, R., et al. 2023, arXiv e-prints, arXiv:2312.14133, doi: 10.48550/arXiv.2312.14133

    work page doi:10.48550/arxiv.2312.14133 2023

  22. [22]

    2021, NewAR, 92, 101595, doi: 10.1016/j.newar.2020.101595

    Ferrigno, C., Savchenko, V., Coleiro, A., et al. 2021, NewAR, 92, 101595, doi: 10.1016/j.newar.2020.101595

    work page doi:10.1016/j.newar.2020.101595 2021

  23. [23]

    Generative Adversarial Networks

    Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., et al. 2014, arXiv e-prints, arXiv:1406.2661, doi: 10.48550/arXiv.1406.2661

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1406.2661 2014

  24. [24]

    Hessels, J. W. T., Spitler, L. G., Seymour, A. D., et al. 2019, ApJL, 876, L23, doi: 10.3847/2041-8213/ab13ae

    work page doi:10.3847/2041-8213/ab13ae 2019

  25. [25]

    M., et al., 2023, @doi [ ] 10.1093/mnras/stad2847 , https://ui.adsabs.harvard.edu/abs/2023MNRAS.526.2039H 526, 2039

    Hewitt, D. M., Hessels, J. W. T., Ould-Boukattine, O. S., et al. 2023, MNRAS, 526, 2039, doi: 10.1093/mnras/stad2847

    work page doi:10.1093/mnras/stad2847 2023

  26. [26]

    Hunter, J. D. 2007, Computing in Science & Engineering, 9, 90, doi: 10.1109/MCSE.2007.55

    work page doi:10.1109/mcse.2007.55 2007

  27. [27]

    D., Burchett, J

    Kilpatrick, C. D., Burchett, J. N., Jones, D. O., et al. 2021, ApJL, 907, L3, doi: 10.3847/2041-8213/abd560

    work page doi:10.3847/2041-8213/abd560 2021

  28. [28]

    D., Tejos, N., Prochaska, J

    Kilpatrick, C. D., Tejos, N., Prochaska, J. X., et al. 2023, arXiv e-prints, arXiv:2311.09316, doi: 10.48550/arXiv.2311.09316

    work page doi:10.48550/arxiv.2311.09316 2023

  29. [29]

    Adam: A Method for Stochastic Optimization

    Kingma, D. P., & Ba, J. 2017, Adam: A Method for Stochastic Optimization. https://arxiv.org/abs/1412.6980

    work page internal anchor Pith review Pith/arXiv arXiv 2017

  30. [30]

    2022, Nature, 602, 585, doi: 10.1038/s41586-021-04354-w

    Kirsten, F., Marcote, B., Nimmo, K., et al. 2022, Nature, 602, 585, doi: 10.1038/s41586-021-04354-w

    work page doi:10.1038/s41586-021-04354-w 2022

  31. [31]

    Koshut, T. M. 1996, PhD thesis, University of Alabama, Huntsville

    work page 1996

  32. [32]

    Kuiper, D., Contardo, G., Huppenkothen, D., & Hessels, J. W. T. 2025, Representation learning for fast radio burst dynamic spectra. https://arxiv.org/abs/2412.12394

    work page arXiv 2025

  33. [33]

    M., Lower, M

    Kumar, P., Shannon, R. M., Lower, M. E., et al. 2022, MNRAS, 512, 3400, doi: 10.1093/mnras/stac683

    work page doi:10.1093/mnras/stac683 2022

  34. [34]

    2019, Journal of Open Source Software, 4, 747, doi: 10.21105/joss.00747

    LeNail, A. 2019, Journal of Open Source Software, 4, 747, doi: 10.21105/joss.00747

    work page doi:10.21105/joss.00747 2019

  35. [35]

    2023a, arXiv e-prints, arXiv:2307.05261, doi: 10.48550/arXiv.2307.05261 —

    Lin, H.-H., Scholz, P., Ng, C., et al. 2023a, arXiv e-prints, arXiv:2307.05261, doi: 10.48550/arXiv.2307.05261 —. 2023b, arXiv e-prints, arXiv:2307.05262, doi: 10.48550/arXiv.2307.05262

    work page doi:10.48550/arxiv.2307.05261

  36. [36]

    R., Bailes, M., McLaughlin, M

    Lorimer, D. R., Bailes, M., McLaughlin, M. A., Narkevic, D. J., & Crawford, F. 2007, Science, 318, 777, doi: 10.1126/science.1147532

    work page doi:10.1126/science.1147532 2007

  37. [37]

    M., & Lee, S.-I

    Lundberg, S. M., & Lee, S.-I. 2017, in Advances in Neural Information Processing Systems, ed. I. Guyon, U. V

    work page 2017

  38. [38]

    https://proceedings.neurips.cc/paper files/paper/2017/ file/8a20a8621978632d76c43dfd28b67767-Paper.pdf

    Associates, Inc.). https://proceedings.neurips.cc/paper files/paper/2017/ file/8a20a8621978632d76c43dfd28b67767-Paper.pdf

    work page 2017

  39. [39]

    2023, MNRAS, 518, 1629, doi: 10.1093/mnras/stac3206

    Luo, J.-W., Zhu-Ge, J.-M., & Zhang, B. 2023, MNRAS, 518, 1629, doi: 10.1093/mnras/stac3206

    work page doi:10.1093/mnras/stac3206 2023

  40. [40]

    J., Men, Y

    Luo, R., Wang, B. J., Men, Y. P., et al. 2020, Nature, 586, 693, doi: 10.1038/s41586-020-2827-2

    work page doi:10.1038/s41586-020-2827-2 2020

  41. [41]

    Marcote, B., Paragi, Z., Hessels, J. W. T., et al. 2017, ApJL, 834, L8, doi: 10.3847/2041-8213/834/2/L8

    work page doi:10.3847/2041-8213/834/2/l8 2017

  42. [42]

    2020, ApJL, 898, L29, doi: 10.3847/2041-8213/aba2cf

    Mereghetti, S., Savchenko, V., Ferrigno, C., et al. 2020, ApJL, 898, L29, doi: 10.3847/2041-8213/aba2cf

    work page doi:10.3847/2041-8213/aba2cf 2020

  43. [43]

    P., Shin, K., et al

    Merryfield, M., Tendulkar, S. P., Shin, K., et al. 2023, AJ, 165, 152, doi: 10.3847/1538-3881/ac9ab5

    work page doi:10.3847/1538-3881/ac9ab5 2023

  44. [44]

    2023, arXiv e-prints, arXiv:2307.01054, doi: 10.48550/arXiv.2307.01054

    Mesarcik, M., Boonstra, A.-J., Iacobelli, M., et al. 2023, arXiv e-prints, arXiv:2307.01054, doi: 10.48550/arXiv.2307.01054

    work page doi:10.48550/arxiv.2307.01054 2023

  45. [45]

    W., et al

    Moroianu, A., Wen, L., James, C. W., et al. 2023, Nature Astronomy, 7, 579, doi: 10.1038/s41550-023-01917-x

    work page doi:10.1038/s41550-023-01917-x 2023

  46. [46]

    2025, The Astronomer’s Telegram, 17081, 1

    Ng, M., & CHIME/FRB Collaboration. 2025, The Astronomer’s Telegram, 17081, 1

    work page 2025

  47. [47]

    Nimmo, K., Hessels, J. W. T., Keimpema, A., et al. 2021, Nature Astronomy, 5, 594, doi: 10.1038/s41550-021-01321-3

    work page doi:10.1038/s41550-021-01321-3 2021

  48. [48]

    Nimmo, K., Hessels, J. W. T., Kirsten, F., et al. 2022, Nature Astronomy, 6, 393, doi: 10.1038/s41550-021-01569-9

    work page doi:10.1038/s41550-021-01569-9 2022

  49. [49]

    R., Marshall, F

    Oates, S. R., Marshall, F. E., Breeveld, A. A., et al. 2021, MNRAS, 507, 1296, doi: 10.1093/mnras/stab2189 O’Malley, T., Bursztein, E., Long, J., et al. 2019, Keras Tuner, https://github.com/keras-team/keras-tuner

    work page doi:10.1093/mnras/stab2189 2021

  50. [50]

    2023, A&A, 678, A149, doi: 10.1051/0004-6361/202243339

    Pastor-Marazuela, I., van Leeuwen, J., Bilous, A., et al. 2023, A&A, 678, A149, doi: 10.1051/0004-6361/202243339

    work page doi:10.1051/0004-6361/202243339 2023

  51. [51]

    2011, Journal of machine learning research, 12, 2825

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, Journal of machine learning research, 12, 2825

    work page 2011

  52. [52]

    2022, arXiv e-prints, arXiv:2210.10615, doi: 10.48550/arXiv.2210.10615

    Peng, Z., Dong, L., Bao, H., Ye, Q., & Wei, F. 2022, arXiv e-prints, arXiv:2210.10615, doi: 10.48550/arXiv.2210.10615

    work page doi:10.48550/arxiv.2210.10615 2022

  53. [53]

    Petroff, E., Hessels, J. W. T., & Lorimer, D. R. 2022, A&A Rv, 30, 2, doi: 10.1007/s00159-022-00139-w

    work page doi:10.1007/s00159-022-00139-w 2022

  54. [54]

    2019, PhR, 821, 1, doi: 10.1016/j.physrep.2019.06.003

    Platts, E., Weltman, A., Walters, A., et al. 2019, PhR, 821, 1, doi: 10.1016/j.physrep.2019.06.003

    work page doi:10.1016/j.physrep.2019.06.003 2019

  55. [55]

    C., Kaspi, V

    Pleunis, Z., Good, D. C., Kaspi, V. M., et al. 2021, ApJ, 923, 1, doi: 10.3847/1538-4357/ac33ac

    work page doi:10.3847/1538-4357/ac33ac 2021

  56. [56]

    L., Haggard, D., et al

    Raza, N., Chan, M. L., Haggard, D., et al. 2023, arXiv e-prints, arXiv:2308.12357, doi: 10.48550/arXiv.2308.12357

    work page doi:10.48550/arxiv.2308.12357 2023

  57. [57]

    R., Curtin, A

    Sand, K. R., Curtin, A. P., Michilli, D., et al. 2025, The Astrophysical Journal, 979, 160, doi: 10.3847/1538-4357/ad9b11

    work page doi:10.3847/1538-4357/ad9b11 2025

  58. [58]

    2020, The Astronomer’s Telegram, 13681, 1 22

    Scholz, P., & CHIME/FRB Collaboration. 2020, The Astronomer’s Telegram, 13681, 1 22

    work page 2020

  59. [59]

    2022, in American Astronomical Society Meeting Abstracts, Vol

    Scholz, P., Kaspi, V., & CHIME/FRB Collaboration. 2022, in American Astronomical Society Meeting Abstracts, Vol. 54, American Astronomical Society Meeting Abstracts, 438.01

    work page 2022

  60. [60]

    G., Cordes, J

    Spitler, L. G., Cordes, J. M., Hessels, J. W. T., et al. 2014, ApJ, 790, 101, doi: 10.1088/0004-637X/790/2/101

    work page doi:10.1088/0004-637x/790/2/101 2014

  61. [61]

    L., Mann, W

    Steinhardt, C. L., Mann, W. J., Rusakov, V., & Jespersen, C. K. 2023, ApJ, 945, 67, doi: 10.3847/1538-4357/acb999

    work page doi:10.3847/1538-4357/acb999 2023

  62. [62]

    P., et al

    Tendulkar, S. P., Bassa, C. G., Cordes, J. M., et al. 2017, ApJL, 834, L7, doi: 10.3847/2041-8213/834/2/L7

    work page doi:10.3847/2041-8213/834/2/l7 2017

  63. [63]

    P., Gil de Paz, A., Kirichenko, A

    Tendulkar, S. P., Gil de Paz, A., Kirichenko, A. Y., et al. 2021, ApJL, 908, L12, doi: 10.3847/2041-8213/abdb38

    work page doi:10.3847/2041-8213/abdb38 2021

  64. [64]

    A., DeLaunay, J., et al

    Tohuvavohu, A., Kennea, J. A., DeLaunay, J., et al. 2020, ApJ, 900, 35, doi: 10.3847/1538-4357/aba94f

    work page doi:10.3847/1538-4357/aba94f 2020

  65. [65]

    2021, ApJ, 915, 102, doi: 10.3847/1538-4357/abfda7

    Verrecchia, F., Casentini, C., Tavani, M., et al. 2021, ApJ, 915, 102, doi: 10.3847/1538-4357/abfda7

    work page doi:10.3847/1538-4357/abfda7 2021

  66. [66]

    Wang, Y.-F., & Nitz, A. H. 2022, ApJ, 937, 89, doi: 10.3847/1538-4357/ac82ae

    work page doi:10.3847/1538-4357/ac82ae 2022

  67. [67]

    R., Chen, P., et al

    Xu, H., Niu, J. R., Chen, P., et al. 2022, Nature, 609, 685, doi: 10.1038/s41586-022-05071-8

    work page doi:10.1038/s41586-022-05071-8 2022

  68. [68]

    2020, PhD thesis, University of British Columbia

    Yadav, P. 2020, PhD thesis, University of British Columbia

    work page 2020

  69. [69]

    B., Wang, J

    Yang, X., Zhang, S. B., Wang, J. S., & Wu, X. F. 2023, MNRAS, 522, 4342, doi: 10.1093/mnras/stad1304

    work page doi:10.1093/mnras/stad1304 2023

  70. [70]

    2022, arXiv e-prints, arXiv:2205.09616, doi: 10.48550/arXiv.2205.09616

    Yi, K., Ge, Y., Li, X., et al. 2022, arXiv e-prints, arXiv:2205.09616, doi: 10.48550/arXiv.2205.09616

    work page doi:10.48550/arxiv.2205.09616 2022

  71. [71]

    2018, arXiv e-prints, arXiv:1809.07294, doi: 10.48550/arXiv.1809.07294

    Yi, X., Walia, E., & Babyn, P. 2018, arXiv e-prints, arXiv:1809.07294, doi: 10.48550/arXiv.1809.07294

    work page doi:10.48550/arxiv.1809.07294 2018

  72. [72]

    2023, MNRAS, 519, 1823, doi: 10.1093/mnras/stac3599

    Zhu-Ge, J.-M., Luo, J.-W., & Zhang, B. 2023, MNRAS, 519, 1823, doi: 10.1093/mnras/stac3599

    work page doi:10.1093/mnras/stac3599 2023