Pith. sign in

REVIEW 3 major objections 6 minor 63 references

Multiband Embeddings of Light Curves

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · 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 a modular ensemble of per-band LSTM networks can classify variable stars from multiband light curves without interpolation or hand-crafted features, and that it beats feature-based random forests in the sparse…

desk verdict Modular multiband RNN with shipped code and real-data results, but the early-LSST claim is contradicted by the paper's own preprocessing and the Gaia comparison fails on macro-recall. read the letter →

arxiv 2501.12499 v2 pith:LO2HII5Z submitted 2025-01-21 astro-ph.IM

I. Becker , P. Protopapas , M. Catelan , K. Pichara This is my paper · ORCID
classification astro-ph.IM
keywords recurrentneuralnetworksmultibandlightcurvesvariablestarclassificationmulti-tasklearningensemblemodelstime-domainastronomyearlyLSST
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper tries to establish that a modular ensemble of recurrent neural networks, one small LSTM per photometric band feeding a central LSTM, can classify variable stars from multiband light curves without computing hand-crafted features, interpolating, or binning the data. Tested on real Gaia, Pan-STARRS1, and ZTF light curves, the ensemble matches or beats Balanced Random Forest classifiers in most classes and shows its largest advantage when light curves are short and split across many bands, the regime expected at the start of LSST. The paper also argues that adding regression of physical parameters such as effective temperature and radius as a secondary task enriches the learned embeddings. The practical payoff, if correct, is a streaming classifier that can update its prediction after every observation with only two model evaluations, making it suitable for alert brokers.

What carries the argument

The mechanism is an ensemble of one LSTM per photometric band, whose hidden states are merged by a learned linear combination, projected through per-band feed-forward translation layers, and fed in observation order into a central LSTM. Inputs are sliding windows of time and magnitude differences with associated uncertainties and an observation-order integer, so the network never interpolates or bins; color enters separately as forward-filled cumulative mean-magnitude differences. Classification is applied at every time step with uncertainty-weighted cross-entropy, and a multi-task branch regresses effective temperature and radius from the final hidden state.

What would settle it

Classify the ZTF test objects in bins of Galactic reddening $E(B-V)$ and check whether the ensemble's macro-F-score falls steadily with extinction while the random forest's does not; if it does, the uncorrected forward-filled colors are not a survey-invariant representation. A sharper test is cross-survey transfer: train on one survey, evaluate on the same physical objects observed by another, and see whether accuracy is preserved.

Watch

Extended reading notes

Core claim

The paper's central claim is that a modular ensemble of RNNs, a dedicated LSTM for each filter plus a central LSTM that fuses their embeddings, is a competitive and more flexible alternative to feature-based classifiers for variable-star light curves. On real data the ensemble reaches median macro-recalls of 0.745 on Gaia, 0.588 on Pan-STARRS1, and 0.828 on ZTF, compared with 0.800, 0.420, and 0.778 for the capped Balanced Random Forest, with per-class F-scores favoring the RNN in most classes and the BRF retaining edges mainly on Gaia T2CEP and ZTF SNIa. The paper argues that the real advantage is in the sparse regime: with few observations per object and many bands, the RNN's F-scores stay high while the random forest's drop because its features cannot be computed reliably. It also claims that training the central network to regress effective temperature and radius alongside classification produces richer embeddings without much harm to classification, and that at prediction time a new observation costs only two model evaluations, one in the relevant single-band LSTM and one in the central LSTM.

Load-bearing premise

The model assumes that colors computed from raw magnitudes, with no reddening correction and with missing band means carried forward, transfer across surveys with different filters; if extinction or filter-system differences shift colors, the central LSTM may learn survey-specific offsets instead of a universal object representation.

Editorial extensions

If this is right

  • Class predictions are available after every observation, so alert brokers could classify sources before full light curves exist, and the model degrades gracefully as observations accumulate.
  • New filters can be added after training by training that band's single-band model and retraining only the central LSTM, without redoing the other bands.
  • At inference, each new observation requires exactly two model evaluations regardless of the number of bands in the survey.
  • The advantage over the Balanced Random Forest grows with more bands and sparser single-band coverage, the configuration expected early in the LSST era.
  • Multi-task regression of effective temperature and radius enriches the learned embedding, and the same embedding can be reused for other regression targets such as orbital parameters of binary systems.

Reading between the lines

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

  • An extension the paper leaves implicit is a formal early-warning policy: because the model outputs a probability at every time step, a broker could issue an alert only when confidence exceeds a threshold and stabilizes over recent steps, instead of at a fixed observation count.
  • The unadjusted forward-filled colors are the fragile link for cross-survey transfer; a natural test is to train with reddening-corrected colors or with extinction and coordinates as additional inputs and see whether the claimed generalization to LSST improves.
  • The learned multiband embedding itself may support transfer learning beyond periodic variables, such as fine-tuning on a small labeled sample of binaries or non-periodic objects, which the paper mentions only as future work.
  • A direct comparison of embeddings of the same physical star observed by two different surveys would reveal whether the central LSTM has learned a survey-invariant representation or has memorized survey-specific color offsets.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The manuscript presents an ensemble of per-band LSTM encoders whose hidden states are projected by learned 'translation layers' and merged by a central LSTM, with multiband colors and time differences appended, for variable-star classification and optional multitask regression of effective temperature and radius. The model is trained and evaluated on three real surveys (Gaia, Pan-STARRS1, ZTF) against a Balanced Random Forest baseline and a concatenation-of-embeddings baseline. The central claims are that the modular architecture adapts to different numbers of bands without interpolation, that it outperforms BRF particularly for sparser multiband light curves, and that its few-observation performance supports early classification for Rubin/LSST.

Significance. The architecture is a reasonable extension of earlier RNN light-curve work, and the paper's empirical per-class F-score tables and released code are valuable. If the headline claims held, a streaming, feature-free multiband classifier would be a practical tool for alert brokers. However, the two central claims are not consistently supported by the reported numbers: on Gaia the proposed Multi model has lower macro-recall than BRF, and no experiment operates in the true first-alert regime of one or two observations per band. The paper itself contains caveats in Sections 5.3 and 6 that conflict with the abstract and conclusion.

major comments (3)
  1. [§5.1–§5.2, §6] The claim in §6 that the proposed model outperforms the BRF in all experiments except Gaia T2CEP and ZTF SNIa is contradicted by the macro-averaged recalls reported in §5.1–§5.2: Gaia Multi=0.745 versus BRF=0.800 and BRF All=0.801, and the macro F-score computed from Table 5 is also not clearly higher for Multi. The text in §5.2 says the model 'generally surpasses' BRF with only two exceptions, which is inaccurate unless Gaia as a whole is exempted. Please reconcile the aggregate metric used (macro recall versus macro F) and either qualify the claim to specific surveys/classes or provide an explanation for the discrepancy.
  2. [§3.1–§3.2, §5.3, §5.4, §6] The abstract and conclusion claim 'superior performance in scenarios with fewer observations, thus providing potential for early classification of sources from facilities such as Vera C. Rubin Observatory's LSST,' but the input representation in §3.1 discards the first observation in each band and, with window w=2, the earliest usable step requires observations 1–3 of that band; §3.2 further excludes the first Nskip predictions (8 for Gaia, 2 for Pan-STARRS1, 3 for ZTF). Section 5.3 explicitly states that the representation 'precludes predictions for the first two observations in each band,' and §6 says the preprocessing 'is not refined for the early classification of transient objects.' Figure 10 begins at 22–25 total observations for Gaia/Pan-STARRS1 and 10 for ZTF, so no experiment tests a genuine first-alert stream. The claims must either be removed or supported by an experiment that reports performance after the first 1–2 observations per band, for example by starting the sequence after the first observation rather than the third.
  3. [§3.4–§3.5, §5.1, Appendix B] Color information, which §5.1 reports as the dominant RF feature (62–84% of Gini importance), is computed without extinction corrections and imputed by forward-filling the cumulative mean magnitude in each band. Because the three surveys use different filter systems and the test objects are not corrected for reddening, the central model may learn survey-specific color offsets rather than a universal object representation; no cross-survey transfer experiment is performed. A concrete test would be leave-one-survey-out training/testing or a comparison with extinction-corrected colors, which would directly address the assumption behind the claimed generalization to LSST.
minor comments (6)
  1. [§5.2] The text describing 'generally surpasses' BRF should report the aggregate macro F-score together with the macro recall to avoid the appearance of cherry-picking; currently §5.1 gives macro recall while Tables 5–7 give per-class F-scores.
  2. [§5.3] Figure 9 is a single-object illustration; a quantitative version showing macro F-score as a function of observation number for the full test set, with the minimum usable observation count marked, would be more informative.
  3. [§1] The same citation 'Bellm et al. (2019)' is used both for ZTF (PASP) and for the LSST Data Management document; please disambiguate the LSST reference (e.g., Ivezić et al. 2019 or the LSST data management white paper).
  4. [Figure 5 caption] The caption 'rows do not sum to one, as the information is derived from different matrices' is unclear for a median confusion matrix; clarify whether the displayed percentages are element-wise medians across the seven folds.
  5. [§3.2] Please justify the chosen Nskip values (8, 2, 3) and state how they were selected; as written, they appear arbitrary and could affect the early-classification comparison.
  6. [Appendix B] The description of forward-fill imputation would benefit from an explicit statement that the carried-forward value is the cumulative mean magnitude of the band in question, and how the first value is initialized for a band before its first observation.

Circularity Check

1 steps flagged · score 1.0 of 10

Minor self-definitional dependency in the auxiliary Teff regression; the main classification claim is benchmarked against external labels and is not circular.

  1. self definitional [Section 2.4 (Physical parameters) and Section 3.4 (Color)]
    "The physical parameters of TIC sources are mainly inferred from Teff, which is obtained from a spline fit as a function of GBP − GRP based on spectroscopic data."

    The regression target Teff is defined as a spline fit of the GBP−GRP color, and the model's input representation explicitly appends colors computed from the same multiband magnitudes (Section 3.4). Therefore, the reported Teff regression (R2=0.86 for Gaia, Table 8) is partly the model relearning the mapping that defined the target, rather than an independent physical validation of the embeddings. The paper itself notes that 'the features correlate strongly with the regressed parameters, which explains the edge of the RF.' This is a minor self-definitional dependency in an auxiliary multitask result; the main classification benchmark uses external class labels and is not affected.

full rationale

The paper's central claim is that an ensemble of RNNs classifies multiband light curves competitively with BRF baselines. This is an empirical benchmark against external catalog class labels, and no classification 'prediction' is a fitted constant of the model, so there is no equation-level circularity. The one mild circularity-adjacent issue is the auxiliary regression of Teff: the TIC target is itself a spline fit of the GBP−GRP color, and the model's input includes exactly this color, so the regression partially recovers the target's defining relation rather than independently validating the embeddings; the paper acknowledges this when it says 'the features correlate strongly with the regressed parameters.' The also-admitted limitation that the input representation precludes predictions for the first two observations per band (§5.3, §6) weakens the abstract's 'early classification for LSST' claim, but this is a scope gap rather than a circular reduction. Overall, the derivation chain is self-contained; the score of 1 reflects the minor regression-target dependency.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The ledger contains no new physical entities. The free parameters are architectural and dataset-construction choices that materially affect the reported scores. The axioms are the labeling assumptions, the use of model-derived regression targets, the color imputation assumption, and the alignment assumption behind the translation layers.

free parameters (7)
  • hidden_state_size = 128 for Gaia and ZTF, 64 for Pan-STARRS1
    Chosen from preliminary experiments rather than fine-tuned (Section 3.2); directly controls model capacity and influences all results.
  • recurrent_layers and hidden-state merge size L = L=3 for all datasets; central LSTM depth 2 for Gaia/ZTF and 3 for Pan-STARRS1
    Architectural hyperparameters set in Sections 3.2 and 3.3; they affect the expressivity of the learned embeddings.
  • Nskip = 8 for Gaia, 2 for Pan-STARRS1, 3 for ZTF
    Early loss terms are skipped to compensate for weak early predictions; this choice shapes the model's behavior on short light curves (Section 3.2).
  • sliding window size and stride = w=2, stride s=1
    Defines the input rows in Equation 2 and determines when the first prediction can be made (Section 3.1).
  • class sampling caps = 40,000 per class for Gaia; 10,000 per class for Pan-STARRS1 and ZTF
    Applied to reduce overfitting and class imbalance (Sections 2.1 to 2.3); changes the training distribution and therefore the macro scores.
  • regression range filters = Teff between 3400 K and 8000 K; radius between 10 and 200 Rsun
    Truncation of target labels in Section 2.4 removes extreme objects and avoids numerical artifacts, affecting the reported regression metrics.
  • lambda_k regression loss weights = not reported
    Equation 6 weights each physical parameter by lambda_k, but the values are never given; the regression training is therefore not fully specified.
assumptions (5)
  • domain assumption Catalog labels from the Gaia DR2 variable catalog and from ALeRCE (Sánchez-Sáez et al. 2021) are correct enough to serve as ground truth for classification.
    Used throughout Section 2 as training and test labels; mislabeled objects would degrade reported F-scores in unknown, class-dependent ways.
  • domain assumption TIC and Apsis physical parameters are valid regression targets, even though they are model-derived estimates.
    Section 2.4 states that TIC Teff comes from a color spline and radius from the Stefan-Boltzmann relation; regression may thus reproduce catalog construction rather than independent physics.
  • domain assumption Colors computed with forward-fill imputation and without extinction correction are informative for classification and do not introduce survey-specific bias.
    Sections 3.4 and 3.5 rely on unreddened, forward-filled colors as inputs to the central model; this is the weakest modeling assumption identified in the review.
  • ad hoc to paper The learned translation layers can map independently trained single-band embeddings into a common representation space.
    Section 3.3 introduces translation layers precisely because single-band embeddings cannot be directly compared; no cross-band training signal enforces alignment before the central model is trained.
  • standard math Standard deep learning assumptions hold: gradient descent finds useful local minima and held-out splits generalize.
    Adopted implicitly throughout Section 3.6; not proven, but standard for this class of empirical work.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Multiband Embeddings of Light Curves." pith.science (2026). https://pith.science/paper/LO2HII5Z

@misc{pith2026250112499,
  author       = {Pith},
  title        = {Pith review of: Multiband Embeddings of Light Curves},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LO2HII5Z}},
  note         = {Machine review of arXiv:2501.12499}
}
read the original abstract

In this work, we propose a novel ensemble of recurrent neural networks (RNNs) that considers the multiband and non-uniform cadence without having to compute complex features. Our proposed model consists of an ensemble of RNNs, which do not require the entire light curve to perform inference, making the inference process simpler. The ensemble is able to adapt to varying numbers of bands, tested on three real light curve datasets, namely Gaia, Pan-STARRS1, and ZTF, to demonstrate its potential for generalization. We also show the capabilities of deep learning to perform not only classification, but also regression of physical parameters such as effective temperature and radius. Our ensemble model demonstrates superior performance in scenarios with fewer observations, thus providing potential for early classification of sources from facilities such as Vera C. Rubin Observatory's LSST. The results underline the model's effectiveness and flexibility, making it a promising tool for future astronomical surveys. Our research has shown that a multitask learning approach can enrich the embeddings obtained by the models, making them instrumental to solve additional tasks, such as determining the orbital parameters of binary systems or estimating parameters for object types beyond periodic ones.

Figures

Figures reproduced from arXiv: 2501.12499 by the authors.

Figure 1
Figure 1. Histograms of the number of observations on each band, per survey. Each band is identified by a color and a hatch style. Gaia filters GBP and GRP are identified with blue solid and orange hatched bars, respectively. Pan-STARRS1 bands are identified with solid green for g, red right diagonal hatch for r, purple horizontal hatch for i, brown left diagonal hatch for z, and pink dotted bar for y. Both Gaia and ZTF are b… view at source ↗
Figure 2
Figure 2. Light curves of the same RRab star with a period of 0.658936 days. The identifiers are 839368005575354240, 169841701110866261 and ZTF17aaajjtn, for Gaia DR2, Pan-STARRS1 DR2 and ZTF DR10, respectively. For Gaia, the observation in the GBP filter is shown in blue circles and orange triangles for GRP. For Pan-STARRS1 on the second panel, the filters gP1,rP1, iP1,zP1, yP1 are identified by green circles, red triangles,… view at source ↗
Figure 3
Figure 3. Single-band representation of a single time step. Solid lines represent the inputs and outputs of the LSTM cells across a single time step. Dashed lines represent the residual connections between recurrent steps. The dotted lines represent the operations to construct the final representation, given in Equation 3. The gray connections represent the inputs and outputs across different time steps [PITH_FULL_IMAGE:figu… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Representation of the multiband ensemble of models. Mr and Mg represent single-band models. Tr and Tg correspond to the translation layers that project the data to a common representation. C is the central model with a similar architecture as shown in [PITH_FULL_IMAGE…
Figure 7
Figure 7. Figure 7: Confusion matrix for ZTF Multi model. The main mistakes are between QSO, AGN and Blazars, given their similar source and a bias towards QSO as one of the most numerous classes [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 6
Figure 6. Figure 6: Confusion matrix for the Pan-STARRS1 Multi model. RRd stars pulsate in the fundamental and first overtone simultaneously, with the latter being the dominant mode most of the time Braga et al. (2022, and references therein). The same behavior can be seen in the fraction…
Figure 8
Figure 8. Figure 8: Distribution of the physical parameter estimations. Each column corresponds to the parameters of a survey [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: presents the classification evolution with re￾spect to the number of multiband observations for a single DSCT_SXPHE object in the Pan-STARRS1 data. Initially, clas￾sification fluctuated between DSCT_SXPHE and RRc classes, while the other classes maintained a low probab…
Figure 10
Figure 10. Figure 10: F-score as a function of the total number of observations. The results of the RNN model are shown in solid blue, while the values for the BRF are shown in diagonal hatched orange. The bins are not uni￾form as some of them were merged to increase the number of objects …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 34 canonical work pages

  1. [1]

    & McEwen, J

    Allam, T. & McEwen, J. D. 2024, RAS Techniques and Instruments, 3, 209

    work page 2024

  2. [2]

    F., Argudo-Fernández, M., et al

    Almeida, A., Anderson, S. F., Argudo-Fernández, M., et al. 2023, ApJS, 267, 44

    2023

  3. [3]

    2018, A&A, 616, A8

    Andrae, R., Fouesneau, M., Creevey, O., et al. 2018, A&A, 616, A8

    2018

  4. [4]

    L., Kiros, J

    Ba, J. L., Kiros, J. R., & Hinton, G. E. 2016, arXiv preprint arXiv:1607.06450

    arXiv 2016

  5. [5]

    2020, MNRAS, 493, 2981

    Becker, I., Pichara, K., Catelan, M., et al. 2020, MNRAS, 493, 2981

    work page 2020

  6. [6]

    2019, Large Synoptic Survey Telescope (LSST) Data Management Article number, page 13 of 15 A&A proofs: manuscript no

    Bellm, E., Blum, R., Graham, M., et al. 2019, Large Synoptic Survey Telescope (LSST) Data Management Article number, page 13 of 15 A&A proofs: manuscript no. one

    work page 2019

  7. [7]

    C., Kulkarni, S

    Bellm, E. C., Kulkarni, S. R., Graham, M. J., et al. 2019, PASP, 131, 018002

    2019

  8. [8]

    B., Ivezi´c, Ž., Jones, R

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

    2022

Show all 63 references
  1. [9]

    F., Fiorentino, G., Bono, G., et al

    Braga, V . F., Fiorentino, G., Bono, G., et al. 2022, MNRAS, 517, 5368

    2022

  2. [10]

    2001, Machine learning, 45, 5

    Breiman, L. 2001, Machine learning, 45, 5

    2001

  3. [11]

    2020, in Advances in Neural Information Processing Systems, ed

    Brown, T., Mann, B., Ryder, N., et al. 2020, in Advances in Neural Information Processing Systems, ed. H. Larochelle, M. Ranzato, R. Hadsell, M. Balcan, & H. Lin, V ol. 33 (Curran Associates, Inc.), 1877–1901

    2020

  4. [12]

    2019, Electronic Imaging, 31, 90

    Brunel, A., Pasquet, J., Pasquet, J., et al. 2019, Electronic Imaging, 31, 90

    2019

  5. [13]

    2023, in Memorie della Societa Astronomica Italiana, V ol

    Catelan, M. 2023, in Memorie della Societa Astronomica Italiana, V ol. 94, 56

    2023

  6. [14]

    & Smith, H

    Catelan, M. & Smith, H. A. 2015, Pulsating Stars (Wiley-VCH)

    2015

  7. [15]

    C., Magnier, E

    Chambers, K. C., Magnier, E. A., Metcalfe, N., et al. 2016, arXiv e-prints, arXiv:1612.05560

    2016 arXiv

  8. [16]

    & Moss, A

    Charnock, T. & Moss, A. 2017, ApJ, 837, L28

    2017

  9. [17]

    2004, University of California, Berkeley, 110, 24

    Chen, C., Liaw, A., Breiman, L., et al. 2004, University of California, Berkeley, 110, 24

    2004

  10. [18]

    2014, arXiv preprint arXiv:1406.1078

    Cho, K., Van Merriënboer, B., Gulcehre, C., et al. 2014, arXiv preprint arXiv:1406.1078

    2014 arXiv

  11. [19]

    2020, arXiv e-prints, arXiv:2009.09796 Dékány, I

    Crawshaw, M. 2020, arXiv e-prints, arXiv:2009.09796 Dékány, I. & Grebel, E. K. 2022, ApJS, 261, 33

    2020 arXiv

  12. [20]

    2018, arXiv e-prints, arXiv:1810.04805

    Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. 2018, arXiv e-prints, arXiv:1810.04805

    2018 arXiv

  13. [21]

    2023, A&A, 670, A54

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

    2023

  14. [22]

    Donoso-Oliva, C., Cabrera-Vives, G., Protopapas, P., Carrasco-Davis, R., & Es- tevez, P. A. 2021, MNRAS, 505, 6069

    2021

  15. [23]

    Fernie, J. D. 1992, AJ, 103, 1647

    1992

  16. [24]

    Fernie, J. D. 1995, AJ, 110, 2361

    1995

  17. [25]

    A., Magnier, E

    Flewelling, H. A., Magnier, E. A., Chambers, K. C., et al. 2020, ApJS, 251, 7 Förster, F., Cabrera-Vives, G., Castillo-Navarrete, E., et al. 2021, AJ, 161, 242 Gaia Collaboration, Vallenari, A., Brown, A. G. A., et al. 2023, A&A, 674, A1

    2020

  18. [26]

    2016, in 2016 IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), 770–778

    He, K., Zhang, X., Ren, S., & Sun, J. 2016, in 2016 IEEE Conference on Com- puter Vision and Pattern Recognition (CVPR), 770–778

    2016

  19. [27]

    & Schmidhuber, J

    Hochreiter, S. & Schmidhuber, J. 1997, Neural Computation, 9, 1735

    1997

  20. [28]

    2018, A&A, 618, A30

    Holl, B., Audard, M., Nienartowicz, K., et al. 2018, A&A, 618, A30

    2018

  21. [29]

    & Szegedy, C

    Ioffe, S. & Szegedy, C. 2015, in Proceedings of the 32nd International Con- ference on International Conference on Machine Learning - V olume 37, ICML’15 (JMLR.org), 448–456 Ivezi´c, Ž., Kahn, S. M., Tyson, J. A., et al. 2019, ApJ, 873, 111 Ivezi´c, Ž., Smith, J. A., Miknaitis...

    2015

  22. [30]

    M., et al

    Jordi, C., Gebran, M., Carrasco, J. M., et al. 2010, A&A, 523, A48

    2010

  23. [31]

    Leavitt, H. S. & Pickering, E. C. 1912, Harvard College Observatory Circular, 173, 1 Lemaître, G., Nogueira, F., & Aridas, C. K. 2017, Journal of Machine Learning Research, 18, 1

    1912

  24. [32]

    2016, in Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (Austin, Texas: Associa- tion for Computational Linguistics), 118–127

    Liu, P., Qiu, X., & Huang, X. 2016, in Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing (Austin, Texas: Associa- tion for Computational Linguistics), 118–127

    2016

  25. [33]

    & Hutter, F

    Loshchilov, I. & Hutter, F. 2019, in International Conference on Learning Rep- resentations

    2019

  26. [34]

    A., Schlafly, E., Finkbeiner, D., et al

    Magnier, E. A., Schlafly, E., Finkbeiner, D., et al. 2013, ApJS, 205, 20

    2013

  27. [35]

    A., Schlafly, E

    Magnier, E. A., Schlafly, E. F., Finkbeiner, D. P., et al. 2020, ApJS, 251, 6

    2020

  28. [36]

    2019, A&A, 621, A144

    Marrese, P., Marinoni, S., Fabrizio, M., & Altavilla, G. 2019, A&A, 621, A144

    2019

  29. [37]

    2021, AJ, 161, 107 Möller, A

    Matheson, T., Stubens, C., Wolf, N., et al. 2021, AJ, 161, 107 Möller, A. & de Boissière, T. 2020, MNRAS, 491, 4277 Möller, A., Peloton, J., Ishida, E. E. O., et al. 2021, MNRAS, 501, 3272

    2021

  30. [38]

    P., & Marshall, J

    Mondrik, N., Long, J. P., & Marshall, J. L. 2015, ApJ, 811, L34

    2015

  31. [39]

    M., Wahl, J

    Morgan, S. M., Wahl, J. N., & Wieckhorst, R. M. 2007, MNRAS, 374, 1421

    2007

  32. [40]

    S., Lochner, M., Webb, S., & Narayan, G

    Muthukrishna, D., Mandel, K. S., Lochner, M., Webb, S., & Narayan, G. 2022, MNRAS, 517, 393

    2022

  33. [41]

    2016, in Proceedings of the 30th International Conference on Neural Information Processing Systems, NIPS’16 (Red Hook, NY , USA: Curran Associates Inc.), 3889–3897

    Neil, D., Pfeiffer, M., & Liu, S.-C. 2016, in Proceedings of the 30th International Conference on Neural Information Processing Systems, NIPS’16 (Red Hook, NY , USA: Curran Associates Inc.), 3889–3897

    2016

  34. [42]

    C., et al

    Ngeow, C.-C., Liao, S.-H., Bellm, E. C., et al. 2021, AJ, 162, 63

    2021

  35. [43]

    2019, A&A, 631, A147

    Nordin, J., Brinnel, V ., van Santen, J., et al. 2019, A&A, 631, A147

    2019

  36. [44]

    2015, arXiv preprint arXiv:1506.00010

    Nun, I., Protopapas, P., Sim, B., et al. 2015, arXiv preprint arXiv:1506.00010

    2015 arXiv

  37. [45]

    A., Wolf, C., Bessell, M

    Onken, C. A., Wolf, C., Bessell, M. S., et al. 2019, PASA, 36, e033

    2019

  38. [46]

    2013, in Proceedings of the 30th In- ternational Conference on International Conference on Machine Learning - V olume 28, ICML’13 (JMLR.org), III–1310–III–1318

    Pascanu, R., Mikolov, T., & Bengio, Y . 2013, in Proceedings of the 30th In- ternational Conference on International Conference on Machine Learning - V olume 28, ICML’13 (JMLR.org), III–1310–III–1318

    2013

  39. [47]

    2019, A&A, 627, A21

    Pasquet, J., Pasquet, J., Chaumont, M., & Fouchez, D. 2019, A&A, 627, A21

    2019

  40. [48]

    2011, Journal of Machine Learning Research, 12, 2825

    Pedregosa, F., Varoquaux, G., Gramfort, A., et al. 2011, Journal of Machine Learning Research, 12, 2825

    2011

  41. [49]

    E., Neumann, M., Iyyer, M., et al

    Peters, M. E., Neumann, M., Iyyer, M., et al. 2018, in Proceedings of the 2018 Conference of the North American Chapter of the Association for Computa- tional Linguistics: Human Language Technologies, V olume 1 (Long Papers) (New Orleans, Louisiana: Association for Computation...

    2018

  42. [50]

    R., Winn, J

    Ricker, G. R., Winn, J. N., Vanderspek, R., et al. 2015, Journal of Astronomical

    2015

  43. [51]

    2021, AJ, 161, 141

    Telescopes, Instruments, and Systems, 1, 014003 Sánchez-Sáez, P., Reyes, I., Valenzuela, C., et al. 2021, AJ, 161, 141

    2021

  44. [52]

    J., Prieto, J

    Shappee, B. J., Prieto, J. L., Grupe, D., et al. 2014, ApJ, 788, 48

    2014

  45. [53]

    W., Williams, R

    Smith, K. W., Williams, R. D., Young, D. R., et al. 2019, RNAAS, 3, 26

    2019

  46. [54]

    Smith, M. J. & Geach, J. E. 2023, Royal Society Open Science, 10, 221454

    2023

  47. [55]

    G., Oelkers, R

    Stassun, K. G., Oelkers, R. J., Paegert, M., et al. 2019, AJ, 158, 138

    2019

  48. [56]

    K., Ackley, K., et al

    Steeghs, D., Galloway, D. K., Ackley, K., et al. 2022, MNRAS, 511, 2405

    2022

  49. [57]

    2015, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    Szegedy, C., Liu, W., Jia, Y ., et al. 2015, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

    2015

  50. [58]

    L., Denneau, L., Heinze, A

    Tonry, J. L., Denneau, L., Heinze, A. N., et al. 2018, PASP, 130, 064505

    2018

  51. [59]

    L., Stubbs, C

    Tonry, J. L., Stubbs, C. W., Lykke, K. R., et al. 2012, ApJ, 750, 99

    2012

  52. [60]

    2023, arXiv preprint arXiv:2307.09288

    Touvron, H., Martin, L., Stone, K., et al. 2023, arXiv preprint arXiv:2307.09288

    2023 arXiv

  53. [61]

    2017, in Advances in Neural Informa- tion Processing Systems, ed

    Vaswani, A., Shazeer, N., Parmar, N., et al. 2017, in Advances in Neural Informa- tion Processing Systems, ed. I. Guyon, U. V . Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, & R. Garnett, V ol. 30 (Curran Associates, Inc.)

    2017

  54. [62]

    2016, arXiv preprint arXiv:1609.08144

    Wu, Y ., Schuster, M., Chen, Z., et al. 2016, arXiv preprint arXiv:1609.08144

    2016 arXiv

  55. [63]

    2016, in Advances in Neural Information Pro- cessing Systems, ed

    Zhang, S., Wu, Y ., Che, T., et al. 2016, in Advances in Neural Information Pro- cessing Systems, ed. D. Lee, M. Sugiyama, U. Luxburg, I. Guyon, & R. Gar- nett, V ol. 29 (Curran Associates, Inc.) Article number, page 14 of 15 I. Becker et al.: Multiband embeddings of light cur...

    2013

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.