pith. sign in

arxiv: 2606.09665 · v1 · pith:EKERZWEBnew · submitted 2026-06-08 · 🌌 astro-ph.GA · astro-ph.IM

A Meta-Learning Framework for Multitask Reverberation Mapping in Active Galactic Nuclei

Aman N. Raju , Andjelka B. Kova\v{c}evi\'c , Dragana Ili\'c , Francesco Tombesi , Luka \v{C}. Popovi\'c , Eric Slezak , Paula Sanchez-Saez , Marina Pavlovi\'c
show 3 more authors
Iva \v{C}vorovi\'c-Hajdinjak Sa\v{s}a Simi\'c {\DJ}or{\dj}e Savi\'c
This is my paper

Pith reviewed 2026-06-27 16:21 UTC · model grok-4.3

classification 🌌 astro-ph.GA astro-ph.IM
keywords active galactic nucleireverberation mappingmeta-learningneural processeslight-curve reconstructiontransfer functionssupermassive black holesphotometric surveys
0
0 comments X

The pith

A meta-learning framework using attentive latent neural processes reconstructs AGN light curves 60-70 percent better than baseline regressors and generalizes from simulations to real data.

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

The paper introduces a meta-learning framework that clusters AGN light curves of similar topology with self-organizing maps and then applies attentive latent neural processes combined with mixture density models to recover light-curve structure, supermassive black hole parameters, and accretion-disk transfer functions without supervision. The approach is tested on simulated light curves that vary in cadence and transfer function, as well as on real Zwicky Transient Facility observations. Relative to ensemble-trained regressors that include Gaussian-process models, the framework produces 60-70 percent better light-curve reconstruction, 35 percent better transfer-function recovery in low-variability clusters, and 34 percent better recovery of intrinsic black-hole and red-noise parameters. The learned latent representations carry information about both transfer functions and black-hole properties, and models trained only on simulations can be applied directly to observed light curves.

Core claim

The attentive latent neural process framework learns latent representations that simultaneously encode transfer functions and supermassive black hole parameters; when the representations are obtained after self-organizing-map clustering and decoded with mixture density models, light-curve reconstruction improves 60-70 percent over ensemble baselines, transfer-function recovery improves approximately 35 percent relative to the training prior in low-variability clusters, and recovery of intrinsic supermassive black hole and red-noise parameters improves approximately 34 percent, with the same models transferring from simulated to real Zwicky Transient Facility light curves.

What carries the argument

Attentive Latent Neural Processes combined with Self-Organizing Maps and Mixture Density Models that perform unsupervised clustering of light-curve topologies and joint recovery of transfer functions and black-hole parameters.

If this is right

  • The method supplies a scalable route to photometric reverberation mapping for the thousands of AGN per square degree expected from wide-field time-domain surveys.
  • Latent representations produced by the framework can be used for joint inference on both transfer functions and black-hole properties without requiring labeled real data.
  • Unsupervised training on simulations followed by direct application to observations removes the need for large annotated training sets drawn from the target survey.
  • The clustering step groups AGN by light-curve topology, which may support population-level studies of variability classes across large samples.

Where Pith is reading between the lines

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

  • The same architecture could be tested on other irregularly sampled time-series problems in astronomy that require recovery of underlying response functions.
  • If the latent space separates transfer-function shape from black-hole mass, it might enable rapid pre-selection of targets for spectroscopic follow-up.
  • Extending the mixture-density output to produce full posterior distributions over transfer functions would allow direct propagation of uncertainty into black-hole mass estimates.
  • The reported gains rest on the particular choice of simulation prior; retraining the same architecture on a different prior could reveal how sensitive the improvements are to simulation realism.

Load-bearing premise

Light-curve topologies and transfer functions generated in the simulated training set match the statistical properties of real AGN light curves observed by current and future surveys closely enough for the unsupervised models to generalize.

What would settle it

A controlled test in which the framework is applied to a large set of real AGN light curves and its reconstruction error exceeds that of the Gaussian-process baseline regressors would falsify the generalization claim.

Figures

Figures reproduced from arXiv: 2606.09665 by Aman N. Raju, Andjelka B. Kova\v{c}evi\'c, {\DJ}or{\dj}e Savi\'c, Dragana Ili\'c, Eric Slezak, Francesco Tombesi, Iva \v{C}vorovi\'c-Hajdinjak, Luka \v{C}. Popovi\'c, Marina Pavlovi\'c, Paula Sanchez-Saez, Sa\v{s}a Simi\'c.

Figure 1
Figure 1. Figure 1: Transfer functions simulated as a sum of a family of relatively displaced Gaussian response functions (top) and corresponding [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Same as Figure [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison of observation strategies. The simulated [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Kernel density distribution of observational data points [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Meta-Learning Framework for multitask reverberation mapping spanning the SOM clustering of light curves, the ALNP [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Benchmark tests of the Meta-Learning Framework assuming moderately (top row) and severely (bottom row) mixed Gaus [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: SOM representation of Cluster 14 (see Fig [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Intrinsic properties of the faster response times (the time lags are in the observer reference frame) in the transfer functions [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Application of the Meta-Learning Framework to di [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Same as Figure [PITH_FULL_IMAGE:figures/full_fig_p016_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Comparison of the reconstruction errors and mean lag distribution for light curves and transfer functions across [PITH_FULL_IMAGE:figures/full_fig_p017_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Comparison of the Neural Processes light curve (blue [PITH_FULL_IMAGE:figures/full_fig_p018_12.png] view at source ↗
Figure 14
Figure 14. Figure 14: Selected cluster (Cluster 7 in Figure C.4) of 20 ZTF g band light curves, with a low variability pattern over a span of 500 days. The gray lines represent the individual light curves, the blue and red lines are the average curve and the SOM rep￾resentation of the cluster curves, respectively. The final portion between 400-500 days is an artifact of padding. Band NLL (log(pm)) MSE(pm2 ) g −1.47 ± 0.14 0.43… view at source ↗
Figure 15
Figure 15. Figure 15: Recovered light curves and transfer functions for object SDSS J002224.18 [PITH_FULL_IMAGE:figures/full_fig_p021_15.png] view at source ↗
read the original abstract

The Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST) is expected to observe active galactic nuclei (AGN) at sky densities of approximately 1000-4000 per sq. deg, enabling photometric reverberation mapping on an unprecedented scale. We present a meta-learning framework for AGN photometric reverberation mapping based on Attentive Latent Neural Processes (ALNP), developed by the SER-SAG-S1 directable software in-kind team for LSST. The framework clusters AGN light curves with similar topologies using Self-Organizing Maps and combines ALNPs with Mixture Density Models to learn light-curve structure, supermassive black hole (SMBH) properties, and accretion-disk transfer functions in an unsupervised manner. We evaluate the framework on simulated AGN light curves spanning a range of cadences and transfer functions, as well as on real data from the Zwicky Transient Facility. The learned latent representations encode information on both transfer functions and SMBH parameters. Relative to ensemble-trained baseline regressors, including Gaussian-process models, the framework improves light-curve reconstruction by 60-70%. The transfer function recovery improves by approximately 35% relative to the training prior in a low-variability cluster, while recovery of intrinsic SMBH and red-noise parameters improves by approximately 34%. We further demonstrate that models trained on simulated data can be applied to real AGN light curves. These results indicate that ALNP-based representations provide a flexible and scalable approach to photometric reverberation mapping and are well suited to the diverse AGN population expected from LSST and future time-domain surveys.

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

3 major / 1 minor

Summary. The manuscript presents a meta-learning framework for photometric reverberation mapping of AGN that combines Attentive Latent Neural Processes (ALNP), Self-Organizing Maps (SOM) to cluster light curves by topology, and Mixture Density Models (MDM). The approach is trained unsupervised on simulated AGN light curves spanning varied cadences and transfer functions, then evaluated on both simulations and real ZTF observations. It reports 60-70% gains in light-curve reconstruction relative to ensemble baselines (including Gaussian processes), ~35% improvement in transfer-function recovery (relative to training prior) within a low-variability cluster, and ~34% improvement in recovery of intrinsic SMBH and red-noise parameters. The work claims that models trained exclusively on simulations can be applied directly to real AGN light curves and positions the method as scalable for LSST-scale surveys.

Significance. If the reported performance gains and sim-to-real generalization are substantiated, the framework would supply a practical unsupervised route to photometric reverberation mapping at the volume expected from LSST (1000-4000 AGN per square degree). The multitask latent representations that jointly encode transfer functions and SMBH parameters, together with the SOM-based clustering, represent a potentially valuable technical advance over single-task regressors for handling heterogeneous AGN populations.

major comments (3)
  1. [Abstract] Abstract: the quantitative claims of 60-70% reconstruction improvement, ~35% transfer-function recovery, and ~34% SMBH/red-noise parameter recovery are stated without specifying the exact metric(s) used, the precise construction of the ensemble-trained baseline regressors (including GP hyperparameter settings and training protocol), error bars, or any statistical significance tests. These omissions are load-bearing because the central claim rests on the magnitude of the reported gains.
  2. [Abstract] Abstract / Results: no quantitative domain-shift diagnostics are supplied between the simulated training distribution and real ZTF light curves (e.g., MMD, KS tests on PSDs or ACFs, or reconstruction error on a real-data hold-out set). Because the framework is trained exclusively on simulations yet applied to real observations, the absence of such checks directly undermines the asserted applicability and the parameter-recovery numbers on real data.
  3. [Evaluation / Methods] Evaluation / Methods: the manuscript does not describe how the ALNP latent encodings are mapped to physical transfer functions and SMBH parameters, nor how the MDM component produces the reported recoveries; without this mapping and without ablation of the SOM clustering step, it is unclear whether the quoted improvements are independent of the simulation priors or simply reflect better interpolation within the training distribution.
minor comments (1)
  1. [Abstract] The abstract would be clearer if it stated the total number of simulated light curves, the exact cadence ranges, and the distribution of transfer-function parameters used in training.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their careful reading and constructive comments. We address each major comment point by point below, indicating the revisions that will be made to the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the quantitative claims of 60-70% reconstruction improvement, ~35% transfer-function recovery, and ~34% SMBH/red-noise parameter recovery are stated without specifying the exact metric(s) used, the precise construction of the ensemble-trained baseline regressors (including GP hyperparameter settings and training protocol), error bars, or any statistical significance tests. These omissions are load-bearing because the central claim rests on the magnitude of the reported gains.

    Authors: We agree that the abstract requires greater specificity to substantiate the quantitative claims. In the revised manuscript we will explicitly state the metrics used for each reported improvement, describe the construction and hyperparameter settings of the ensemble baselines (including Gaussian processes), and report error bars together with the results of statistical significance tests. revision: yes

  2. Referee: [Abstract] Abstract / Results: no quantitative domain-shift diagnostics are supplied between the simulated training distribution and real ZTF light curves (e.g., MMD, KS tests on PSDs or ACFs, or reconstruction error on a real-data hold-out set). Because the framework is trained exclusively on simulations yet applied to real observations, the absence of such checks directly undermines the asserted applicability and the parameter-recovery numbers on real data.

    Authors: We acknowledge that quantitative domain-shift diagnostics are necessary to support the sim-to-real claims. The revised manuscript will include such diagnostics, for example MMD or KS tests on PSDs and ACFs between the simulated and ZTF distributions, as well as reconstruction performance on any available real-data hold-out sets. revision: yes

  3. Referee: [Evaluation / Methods] Evaluation / Methods: the manuscript does not describe how the ALNP latent encodings are mapped to physical transfer functions and SMBH parameters, nor how the MDM component produces the reported recoveries; without this mapping and without ablation of the SOM clustering step, it is unclear whether the quoted improvements are independent of the simulation priors or simply reflect better interpolation within the training distribution.

    Authors: We thank the referee for identifying this omission. The revised Methods and Evaluation sections will describe the mapping from ALNP latent encodings to physical transfer functions and SMBH parameters via the MDM, and will include an ablation study of the SOM clustering step to clarify the origin of the reported gains. revision: yes

Circularity Check

0 steps flagged

No significant circularity; framework evaluated on held-out simulated and separate real data

full rationale

The paper describes a meta-learning framework (ALNP + SOM + MDM) trained on simulated AGN light curves and applied to both held-out simulations and independent real ZTF observations. Performance gains (60-70% reconstruction, 34-35% parameter recovery) are reported relative to external ensemble baselines including Gaussian processes. No equations or steps reduce a claimed prediction to a fitted input by construction, and no load-bearing self-citation chain is invoked to justify uniqueness or ansatz choices. The derivation remains self-contained against the stated external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available, so free parameters, axioms, and invented entities cannot be enumerated. The neural network components likely contain many fitted hyperparameters whose values are not reported.

pith-pipeline@v0.9.1-grok · 5904 in / 1166 out tokens · 23757 ms · 2026-06-27T16:21:28.359712+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

106 extracted references · 7 canonical work pages · 2 internal anchors

  1. [1]

    2015, Information Systems, 53, 16 Arévalo, P., Churazov, E., Lira, P., et al

    Aghabozorgi, S., Seyed Shirkhorshidi, A., & Ying Wah, T. 2015, Information Systems, 53, 16 Arévalo, P., Churazov, E., Lira, P., et al. 2024, A&A, 684, A133 Arévalo, P., Churazov, E., Zhuravleva, I., Hernández-Monteagudo, C., &

    2015

  2. [2]

    2012, MNRAS, 426, 1793 Arévalo, P., Lira, P., Sánchez-Sáez, P., et al

    Revnivtsev, M. 2012, MNRAS, 426, 1793 Arévalo, P., Lira, P., Sánchez-Sáez, P., et al. 2023, MNRAS, 526, 6078

    2012

  3. [3]

    J., Temple, M., Richards, G., Yu, W., & Bauer, F

    Assef, R. J., Temple, M., Richards, G., Yu, W., & Bauer, F. 2021, LSST AGN SC Cadence Note: Quasar Counts, accessed: 2025-02-18

    2021

  4. [4]

    2017, in 2017 IEEE International Conference on Computer Vision (ICCV), 3306–3314

    Bahat, Y ., Efrat, N., & Irani, M. 2017, in 2017 IEEE International Conference on Computer Vision (ICCV), 3306–3314

    2017

  5. [5]

    Bardeen, J. M. & Petterson, J. A. 1975, ApJ, 195, L65

    1975

  6. [6]

    & Kowalski, M

    Bartos, I. & Kowalski, M. 2017, Multimessenger Astronomy

    2017

  7. [7]

    2009, ApJ, 696, 1241

    Bauer, A., Baltay, C., Coppi, P., et al. 2009, ApJ, 696, 1241

    2009

  8. [8]

    Bellm, E. C. e. a. 2019, PASP, 131, 018002

    2019

  9. [9]

    B., Ivezi´c, Ž., Jones, R

    Bianco, F. B., Ivezi´c, Ž., Jones, R. L., et al. 2022, ApJS, 258, 1

    2022

  10. [10]

    Bishop, C. M. 2006, Pattern Recognition and Machine Learning, Information Science and Statistics (Berlin: Springer)

    2006

  11. [11]

    Blandford, R. D. & McKee, C. F. 1982, ApJ, 255, 419

    1982

  12. [12]

    2023, The Astrophysical Journal Sup- plement Series, 265, 27

    Bonito, R., Venuti, L., Ustamujic, S., et al. 2023, The Astrophysical Journal Sup- plement Series, 265, 27

    2023

  13. [13]

    J., Ford, K

    Breivik, K., Connolly, A. J., Ford, K. E. S., et al. 2022, arXiv e-prints, arXiv:2208.02781

    work page arXiv 2022

  14. [14]

    R., West, R

    Brett, D. R., West, R. G., & Wheatley, P. J. 2004, Monthly Notices of the Royal Astronomical Society, 353, 369

    2004

  15. [15]

    J., Shen, Y ., Blaes, O., et al

    Burke, C. J., Shen, Y ., Blaes, O., et al. 2021, Science, 373, 789

    2021

  16. [16]

    M., Bentz, M

    Cackett, E. M., Bentz, M. C., & Kara, E. 2021, iScience, 24, 102557

    2021

  17. [17]

    M., Horne, K., & Winkler, H

    Cackett, E. M., Horne, K., & Winkler, H. 2007, MNRAS, 380, 669

    2007

  18. [18]

    2017, ApJ, 834, 111

    Caplar, N., Lilly, S., & Trakhtenbrot, B. 2017, ApJ, 834, 111

    2017

  19. [19]

    Chan, J. H. H., Millon, M., Bonvin, V ., & Courbin, F. 2020, A&A, 636, A52

    2020

  20. [20]

    2022, Monthly Notices of the Royal As- tronomical Society: Letters, 511, 13

    Chand, K., Omar, A., Chand, H., et al. 2022, Monthly Notices of the Royal As- tronomical Society: Letters, 511, 13

    2022

  21. [21]

    & Daniel, E

    Chelouche, D. & Daniel, E. 2012, ApJ, 747, 62

    2012

  22. [22]

    & Wang, J

    Chen, X. & Wang, J. 2015, ApJ, 805, 80 De Cicco, D., Bauer, F. E., Paolillo, M., et al. 2021, A&A, 645, A103 De Cicco, D., Paolillo, M., Falocco, S., et al. 2019, A&A, 627, A33 de Vries, W. H., Becker, R. H., White, R. L., & Loomis, C. 2005, AJ, 129, 615

    2015

  23. [23]

    2025, ApJ, 980, 257

    Deesamutara, S., Chainakun, P., Worrakitpoonpon, T., et al. 2025, ApJ, 980, 257

    2025

  24. [24]

    G., Drake, A

    Djorgovski, S. G., Drake, A. J., Mahabal, A. A., et al. 2016, in American Astro- nomical Society Meeting Abstracts, V ol. 227, American Astronomical Soci- ety Meeting Abstracts #227, 349.14

    2016

  25. [25]

    2023, A&A, 670, A54

    Donoso-Oliva, C., Becker, I., Protopapas, P., et al. 2023, A&A, 670, A54

    2023

  26. [26]

    J., Djorgovski, S

    Drake, A. J., Djorgovski, S. G., Catelan, M., et al. 2017, Monthly Notices of the Royal Astronomical Society, 469, 3688

    2017

  27. [27]

    Dubois, Y ., Gordon, J., & Foong, A. Y . K. 2020, Neural Process Family,http: //yanndubs.github.io/Neural-Process-Family/

    2020

  28. [28]

    C., & Trump, J

    Elitzur, M., Ho, L. C., & Trump, J. R. 2014, MNRAS, 438, 3340 Event Horizon Telescope Collaboration, Akiyama, K., Alberdi, A., et al. 2022, ApJ, 930, L12 Event Horizon Telescope Collaboration, Akiyama, K., Alberdi, A., et al. 2019, ApJ, 875, L4

    2014

  29. [29]

    2024a, arXiv e-prints, arXiv:2410.18423

    Fagin, J., Hung-Hsu Chan, J., Best, H., et al. 2024a, arXiv e-prints, arXiv:2410.18423

    work page arXiv

  30. [30]

    D., Bianco, F

    Feigelson, E. D., Bianco, F. B., & Bonito, R. 2023, The Astrophysical Journal Supplement Series, 268, 11

    2023

  31. [31]

    Foong, A. Y . K., Bruinsma, W. P., Gordon, J., et al. 2020, in Proceedings of the 34th International Conference on Neural Information Processing Systems, NIPS ’20 (Red Hook, NY , USA: Curran Associates Inc.)

    2020

  32. [32]

    2018, in Proceedings of Ma- chine Learning Research, V ol

    Garnelo, M., Rosenbaum, D., Maddison, C., et al. 2018, in Proceedings of Ma- chine Learning Research, V ol. 80, Proceedings of the 35th International Con- ference on Machine Learning, ed. J. Dy & A. Krause (PMLR), 1704–1713

    2018

  33. [33]

    Neural Processes

    Garnelo, M., Schwarz, J., Rosenbaum, D., et al. 2018, arXiv e-prints, arXiv:1807.01622

    work page internal anchor Pith review Pith/arXiv arXiv 2018

  34. [34]

    1999, MNRAS, 306, 637

    Giveon, U., Maoz, D., Kaspi, S., Netzer, H., & Smith, P. 1999, MNRAS, 306, 637

    1999

  35. [35]

    2016, Deep Learning (MIT Press)

    Goodfellow, I., Bengio, Y ., & Courville, A. 2016, Deep Learning (MIT Press)

    2016

  36. [36]

    2019, Publications of the Astronomi- cal Society of the Pacific, 131, 078001

    Graham, M., Kulkarni, S., Bellm, E., et al. 2019, Publications of the Astronomi- cal Society of the Pacific, 131, 078001

    2019

  37. [37]

    J., Djorgovski, S

    Graham, M. J., Djorgovski, S. G., Stern, D., et al. 2015, MNRAS, 453, 1562

    2015

  38. [38]

    Griffiths, R.-R., Jiang, J., Buisson, D. J. K., et al. 2021, ApJ, 914, 144 Guérin, A., Chauvet, P., & Saubion, F. 2024, arXiv e-prints, arXiv:2501.08416

    work page arXiv 2021

  39. [39]

    2002, Monthly Notices of the Royal Astronomical Society, 329, 76

    Hawkins, M. 2002, Monthly Notices of the Royal Astronomical Society, 329, 76

    2002

  40. [40]

    F., Coughlin, M

    Healy, B. F., Coughlin, M. W., Mahabal, A. A., et al. 2024, ApJS, 272, 14

    2024

  41. [41]

    I., Harrison, T

    Hoffman, D. I., Harrison, T. E., & McNamara, B. J. 2009, The Astronomical Journal, 138, 466

    2009

  42. [42]

    R., Grier, C

    Homayouni, Y ., Trump, J. R., Grier, C. J., et al. 2019, ApJ, 880, 126 Hönig, S. F. 2014, ApJ, 784, L4

    2019

  43. [43]

    M., Collier, S

    Horne, K., Peterson, B. M., Collier, S. J., & Netzer, H. 2004, PASP, 116, 465

    2004

  44. [44]

    2022, IEEE Transac- tions on Pattern Analysis and Machine Intelligence, 44, 5149

    Hospedales, T., Antoniou, A., Micaelli, P., & Storkey, A. 2022, IEEE Transac- tions on Pattern Analysis and Machine Intelligence, 44, 5149

    2022

  45. [45]

    2019, MNRAS, 488, 324 Ivezi´c, Ž., Kahn, S

    Ingram, A., Mastroserio, G., Dauser, T., et al. 2019, MNRAS, 488, 324 Ivezi´c, Ž., Kahn, S. M., Tyson, J. A., et al. 2019, ApJ, 873, 111

    2019

  46. [46]

    K., Joshi, R., Chand, H., et al

    Jha, V . K., Joshi, R., Chand, H., et al. 2022, Monthly Notices of the Royal As- tronomical Society, 511, 3005

    2022

  47. [47]

    P., V ogeley, M

    Kasliwal, V . P., V ogeley, M. S., & Richards, G. T. 2017, MNRAS, 470, 3027

    2017

  48. [48]

    P., V ogeley, M

    Kasliwal, V . P., V ogeley, M. S., Richards, G. T., Williams, J., & Carini, M. T. 2015, MNRAS, 453, 2075

    2015

  49. [49]

    C., Bechtold, J., & Siemiginowska, A

    Kelly, B. C., Bechtold, J., & Siemiginowska, A. 2009, ApJ, 698, 895

    2009

  50. [50]

    C., Becker, A

    Kelly, B. C., Becker, A. C., Sobolewska, M., Siemiginowska, A., & Uttley, P. 2014, ApJ, 788, 33

    2014

  51. [51]

    2019, in International Conference on Learning Representations

    Kim, H., Mnih, A., Schwarz, J., et al. 2019, in International Conference on Learning Representations

    2019

  52. [52]

    Kingma, D. P. & Ba, J. 2014, arXiv e-prints, arXiv:1412.6980

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  53. [53]

    S., Shappee, B

    Kochanek, C. S., Shappee, B. J., Stanek, K. Z., et al. 2017, Publications of the Astronomical Society of the Pacific, 129, 104502

    2017

  54. [54]

    1990, Proceedings of the IEEE, 78, 1464

    Kohonen, T. 1990, Proceedings of the IEEE, 78, 1464

    1990

  55. [55]

    F., Blanc, G

    Kollmeier, J., Anderson, S. F., Blanc, G. A., et al. 2019, in Bulletin of the Amer- ican Astronomical Society, V ol. 51, 274

    2019

  56. [56]

    S., Sijacki, D., et al

    Koudmani, S., Somerville, R. S., Sijacki, D., et al. 2024, Monthly Notices of the Royal Astronomical Society, 532, 60 Kovaˇcevi´c, A. B., Ili ´c, D., Popovi ´c, L. ˇC., et al. 2021, Monthly Notices of the Royal Astronomical Society, 505, 5012 Kovaˇcevi´c, A. B., Ili´c, D., Popovi´c, L., et al. 2023, Universe, 9, 287 Kovaˇcevi´c, A. B., Radovi´c, V ., Ili´...

    2024

  57. [57]

    2020, MNRAS, 499, 6053

    Laurenti, M., Vagnetti, F., Middei, R., & Paolillo, M. 2020, MNRAS, 499, 6053

    2020

  58. [58]

    M., Kulkarni, S

    Law, N. M., Kulkarni, S. R., Dekany, R. G., et al. 2009, Publications of the Astronomical Society of the Pacific, 121, 1395

    2009

  59. [59]

    I., & Bianco, F

    Li, X., Ragosta, F., Clarkson, W. I., & Bianco, F. B. 2021, The Astrophysical Journal Supplement Series, 258, 2

    2021

  60. [60]

    2016, ApJ, 831, 206

    Li, Y .-R., Wang, J.-M., & Bai, J.-M. 2016, ApJ, 831, 206

    2016

  61. [61]

    2016, The Astrophysical Journal, 831, 206

    Li, Y .-R., Wang, J.-M., & Bai, J.-M. 2016, The Astrophysical Journal, 831, 206

    2016

  62. [62]

    2018, ApJ, 861, 6

    Li, Z., McGreer, I., Wu, X., Fan, X., & Yang, Q. 2018, ApJ, 861, 6

    2018

  63. [63]

    Y .-Y ., Pandya, S., Pratap, D., et al

    Lin, J. Y .-Y ., Pandya, S., Pratap, D., et al. 2022, Monthly Notices of the Royal Astronomical Society, 518, 4921

    2022

  64. [64]

    2020, MNRAS, 494, 3686

    Luo, Y ., Shen, Y ., & Yang, Q. 2020, MNRAS, 494, 3686

    2020

  65. [65]

    2012, ApJ, 753, 106

    MacLeod, C., Ivezi´c, Ž., Sesar, B., et al. 2012, ApJ, 753, 106

    2012

  66. [66]

    L., Ivezi´c, Ž., Kochanek, C

    MacLeod, C. L., Ivezi´c, Ž., Kochanek, C. S., et al. 2010, ApJ, 721, 1014

    2010

  67. [67]

    2019, Publications of the As- tronomical Society of the Pacific, 131, 038002

    Mahabal, A., Rebbapragada, U., Walters, R., et al. 2019, Publications of the As- tronomical Society of the Pacific, 131, 038002

    2019

  68. [68]

    J., Laher, R

    Masci, F. J., Laher, R. R., Rusholme, B., et al. 2019, PASP, 131, 018003

    2019

  69. [69]

    & Basford, K

    Mclachlan, G. & Basford, K. 1988, Mixture Models: Inference and Applications to Clustering, V ol. 38 Article number, page 23 of 34 A&A proofs:manuscript no. main

    1988

  70. [70]

    J., Xiang, Z., & Kesidis, G

    Miller, D. J., Xiang, Z., & Kesidis, G. 2023, Adversarial Learning and Secure AI (Cambridge University Press)

    2023

  71. [71]

    S., Richards, G

    Moreno, J., V ogeley, M. S., Richards, G. T., & Yu, W. 2019, PASP, 131, 063001

    2019

  72. [72]

    W., Hyer, G

    Morgan, C. W., Hyer, G. E., Bonvin, V ., et al. 2018, ApJ, 869, 106

    2018

  73. [73]

    W., Kochanek, C

    Morgan, C. W., Kochanek, C. S., Morgan, N. D., & Falco, E. E. 2010, The As- trophysical Journal, 712, 1129

    2010

  74. [74]

    S., Chambers, K

    Morganson, E., Burgett, W. S., Chambers, K. C., et al. 2014, ApJ, 784, 92

    2014

  75. [75]

    F., Edelson, R., Baumgartner, W., & Gandhi, P

    Mushotzky, R. F., Edelson, R., Baumgartner, W., & Gandhi, P. 2011, ApJ, 743, L12

    2011

  76. [76]

    K., Chand, H., & Singh, V

    Ojha, V ., Jha, V . K., Chand, H., & Singh, V . 2022, Monthly Notices of the Royal Astronomical Society, 514, 5607

    2022

  77. [77]

    M., Assef, R

    Padovani, P., Alexander, D. M., Assef, R. J., et al. 2017, A&A Rev., 25, 2

    2017

  78. [78]

    Panaretos, V . M. & Zemel, Y . 2019, Annual Review of Statistics and Its Appli- cation, 6, 405

    2019

  79. [79]

    W., Villar, A., Li, Y ., et al

    Park, J. W., Villar, A., Li, Y ., et al. 2021, arXiv e-prints, arXiv:2106.01450

    work page arXiv 2021

  80. [80]

    Peterson, B. M. & Horne, K. 2004, Astronomische Nachrichten, 325, 248

    2004

Showing first 80 references.