pith. sign in

arxiv: 2605.23700 · v1 · pith:KJCPEN3Rnew · submitted 2026-05-22 · 🌌 astro-ph.EP · astro-ph.IM· astro-ph.SR

New substellar candidates identified through deep learning in the F150 sample of the large-scale SHINE direct imaging survey

Carles Cantero Mitjans , Mariam Sabalbal , Olivier Absil , Marc Van Droogenbroeck , Damien S\'egransan , Philippe Delorme This is my paper

Pith reviewed 2026-05-25 02:35 UTC · model grok-4.3

classification 🌌 astro-ph.EP astro-ph.IMastro-ph.SR
keywords direct imagingexoplanetsbrown dwarfsdeep learningSHINE surveysubstellar candidatesangular differential imaging
0
0 comments X

The pith

Deep learning on SHINE F150 data recovers all known companions and flags 13 new substellar candidates

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

The paper applies the NA-SODINN supervised deep learning model to reprocess the F150 sample of 150 stars from the SHINE direct imaging survey. The model recovers every previously known companion and several debris disks while also surfacing 13 previously unreported substellar candidates. Ten of the new candidates appear in both H2 and H3 bands; color-magnitude analysis marks three as photometrically promising, yet only the source near Smethells 20 survives as a strong follow-up target when checked against existing multi-epoch SPHERE observations.

Core claim

NA-SODINN recovers all known companions and some debris disks in the F150 sample and identifies 13 new substellar candidates not reported in previous studies: ten detected in both the H2 and H3 bands, and three in only one band. For the ten sources detected in both bands, H2-H3 color-magnitude analysis identifies three photometrically promising candidates, but only the candidate around Smethells 20 remains a strong target for follow-up given current multi-epoch SPHERE data.

What carries the argument

NA-SODINN, a supervised deep learning model that produces pixel-wise confidence maps from angular differential imaging sequences, paired with an F1-score-based thresholding strategy to convert maps into detections while balancing sensitivity and specificity.

If this is right

  • The same NA-SODINN pipeline can be run on the remaining SHINE targets to search for additional overlooked companions.
  • Color-magnitude diagrams provide an initial filter that narrows candidate lists before expensive multi-epoch observations.
  • Only sources with supporting multi-epoch astrometry advance to high-priority follow-up status.
  • The method demonstrates that data-driven detection can complement classical speckle-subtraction techniques without requiring new telescope time.

Where Pith is reading between the lines

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

  • Reprocessing existing large surveys with modern models could increase the total yield of directly imaged substellar objects without new observations.
  • If the Smethells 20 candidate is confirmed, it would add one more benchmark object for testing formation models at wide separations.
  • The F1-score thresholding approach may generalize to other high-contrast imaging instruments facing similar speckle-noise problems.

Load-bearing premise

The F1-score-based thresholding strategy applied to the model's pixel-wise confidence maps produces reliable detections that accurately distinguish real companions from residual speckle noise without introducing a large number of false positives.

What would settle it

High-contrast follow-up imaging or spectroscopy that either confirms or rules out the candidate around Smethells 20 as a bound substellar companion.

Figures

Figures reproduced from arXiv: 2605.23700 by Carles Cantero Mitjans, Damien S\'egransan, Marc Van Droogenbroeck, Mariam Sabalbal, Olivier Absil, Philippe Delorme.

Figure 1
Figure 1. Figure 1: Distributions of the main features extracted from the ADI sequences of the F150 sample. The number of frames in the [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of the training set used by NA-SODINN, showing three examples of patch sequences from the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Comparison between the S/N map (left), computed using annular PCA with k = 20 principal components, and the confidence map (right), computed using NA-SODINN trained with faint companion injections, for the sph1 SPHERE ADI sequence from EIDC (Cantalloube et al. 2020). The small dashed circles in white indicate the true locations of the injected fake companions c1 and c2, while the red circles mark the posit… view at source ↗
Figure 4
Figure 4. Figure 4: Output of NA-SODINN’s proposed thresholding strategy for the [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: H2–H3 color–magnitude diagram of the 10 NA-SODINN [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: NA-SODINN contrast curves, based on 95% complete [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
read the original abstract

Context. The SPHERE High-contrast Imaging survey for Exoplanets (SHINE) represents one of the largest direct imaging campaigns, targeting over 400 young, nearby stars with the goal of detecting and characterizing giant exoplanets and brown dwarfs. This dataset offers a unique opportunity to revisit observations using modern, data-driven approaches, potentially uncovering new substellar candidates that may have been overlooked by classical analysis techniques. Aims. Our study focuses on reprocessing and reanalyzing the so-called F150 sample, a well-defined subset of 150 main-sequence stars within 100 pc observed in the H-band with VLT/SPHERE as part of the SHINE survey. Methods. We apply NA-SODINN, a supervised deep learning model specifically tailored for detecting faint planetary signals in angular differential imaging (ADI) sequences. Designed to model local noise properties and capture spatial context, NA-SODINN is effective at distinguishing real companions from residual speckle noise. To translate the model's pixel-wise confidence maps into actionable detections, we introduce a novel F1-score-based thresholding strategy. This principled approach balances sensitivity and specificity, addressing a key limitation in current deep learning-based methods. Results. NA-SODINN recovers all known companions and some of the debris disks in the F150 sample, and identifies 13 new substellar candidates not reported in previous studies: ten detected in both the H2 and H3 bands, and three in only one band. For the ten sources detected in both bands, we use the H2-H3 color-magnitude diagram to perform a first assessment of their nature. Based on this analysis, we identify two ambiguous cases and three photometrically promising candidates. However, in light of the currently available multi-epoch SPHERE data, only the candidate around Smethells 20 remains a strong target for follow-up.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript applies the NA-SODINN supervised deep learning model to the F150 sample (150 stars) of the SHINE VLT/SPHERE survey. It reports recovery of all previously known companions plus some debris disks, and claims the detection of 13 new substellar candidates (10 in both H2/H3 bands, 3 in one band). Color-magnitude analysis of the dual-band sources identifies three photometrically promising candidates, with only the source near Smethells 20 retained as a strong follow-up target given existing multi-epoch data. A novel F1-score-based thresholding procedure is introduced to convert model confidence maps into detections.

Significance. If the new candidates are robust, the work would illustrate that modern data-driven techniques can extract additional companions from existing large direct-imaging surveys that classical pipelines may have missed. Recovery of all known signals supplies a basic sensitivity check, and the F1-thresholding method addresses a recognized gap in DL-based detection pipelines. The photometric triage of the dual-band sources is a useful first step toward prioritization.

major comments (2)
  1. [Methods (thresholding strategy)] The central claim of 13 new substellar candidates rests on the F1-score-based thresholding applied to NA-SODINN confidence maps (Methods section describing the thresholding strategy). The manuscript does not report quantitative false-positive rates, results from negative-control injections, or tests of threshold stability across the heterogeneous noise properties of the 150 targets; recovery of known companions tests sensitivity but supplies no direct evidence on specificity for the new detections.
  2. [Results (color-magnitude analysis)] The photometric assessment of the ten dual-band candidates (color-magnitude diagram analysis) identifies three promising sources, yet the text provides no error bars on the H2-H3 colors, no comparison to field or young-object sequences with uncertainties, and no quantitative assessment of how residual speckle or background contamination could affect the classification.
minor comments (2)
  1. [Abstract] The abstract states that 'all known companions are recovered' but does not specify how many such companions exist in the F150 sample or whether any were missed at low signal-to-noise.
  2. [Methods] Notation for the NA-SODINN model and the F1-thresholding procedure should be defined more explicitly (e.g., what constitutes a 'detection' after thresholding) to allow reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive review and the opportunity to address these points. We respond to each major comment below and will revise the manuscript accordingly.

read point-by-point responses

  1. Referee: [Methods (thresholding strategy)] The central claim of 13 new substellar candidates rests on the F1-score-based thresholding applied to NA-SODINN confidence maps (Methods section describing the thresholding strategy). The manuscript does not report quantitative false-positive rates, results from negative-control injections, or tests of threshold stability across the heterogeneous noise properties of the 150 targets; recovery of known companions tests sensitivity but supplies no direct evidence on specificity for the new detections.

    Authors: We agree that the current validation emphasizes sensitivity via recovery of known companions but does not directly quantify specificity for the new candidates. In the revised manuscript we will add false-positive rate estimates obtained from negative-control injections (phase-shuffled ADI cubes and synthetic noise realizations) and will report the stability of the chosen F1-score threshold on a representative subset of targets spanning the observed range of noise properties. These additions will be placed in the Methods section alongside the existing F1-thresholding description. revision: yes

  2. Referee: [Results (color-magnitude analysis)] The photometric assessment of the ten dual-band candidates (color-magnitude diagram analysis) identifies three promising sources, yet the text provides no error bars on the H2-H3 colors, no comparison to field or young-object sequences with uncertainties, and no quantitative assessment of how residual speckle or background contamination could affect the classification.

    Authors: We acknowledge these omissions in the photometric triage. The revised version will include (i) error bars on the measured H2–H3 colors derived from the NA-SODINN confidence maps and aperture photometry, (ii) direct overlays of the field and young-object sequences with their published uncertainties, and (iii) a quantitative discussion of possible contamination by residual speckles or background sources, including an estimate of the expected number of false positives based on the local noise statistics. These elements will be added to the Results section describing the color-magnitude analysis. revision: yes

Circularity Check

0 steps flagged

NA-SODINN F1-thresholding and detections remain independent of the reported new candidates

full rationale

The paper applies a supervised deep-learning model (NA-SODINN) trained on simulated or external signals to the F150 observational dataset, then converts pixel-wise confidence maps to detections via an F1-score-based threshold. Recovery of all known companions serves as a sensitivity check on the same data, but the threshold choice and model weights are not defined in terms of the 13 new candidates; the color-magnitude assessment of those candidates is performed after detection and does not feed back into the model or threshold. No equations, self-citations, or ansatzes reduce the headline result to a quantity fitted from the identical inputs by construction. This is the normal case of an applied ML pipeline evaluated on held-out astronomical data.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

No free parameters, invented entities, or non-standard axioms are described in the abstract. The work rests on standard domain assumptions of angular differential imaging and supervised learning on simulated companions.

axioms (1)
  • domain assumption Angular differential imaging sequences allow separation of companion signals from quasi-static speckle noise under the assumptions of the SHINE observing strategy.
    The method is applied to ADI sequences from the survey.

pith-pipeline@v0.9.0 · 5908 in / 1344 out tokens · 26103 ms · 2026-05-25T02:35:56.303931+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

144 extracted references · 144 canonical work pages

  1. [1]

    2016, in USENIX Conf

    Abadi, M., Barham, P., Chen, J., et al. 2016, in USENIX Conf. Oper. Syst. Des. Implement. (OSDI), 265–283

    work page 2016

  2. [2]

    2013, , 559, 1

    Absil, O., Milli, J., Mawet, D., et al. 2013, , 559, 1

    work page 2013

  3. [3]

    & Quanz, S

    Amara, A. & Quanz, S. P. 2012, , 427, 948

    work page 2012

  4. [4]

    & Tuomi, M

    Anglada-Escudé, G. & Tuomi, M. 2012, , 548, A58

    work page 2012

  5. [5]

    Bell , C. P. M., Mamajek , E. E., & Naylor , T. 2015, , 454, 593

    work page 2015

  6. [6]

    2019, , 631, 1

    Beuzit, J.-L., Vigan, A., Mouillet, D., et al. 2019, , 631, 1

    work page 2019

  7. [7]

    A., Liu, M

    Biller, B. A., Liu, M. C., Wahhaj, Z., et al. 2010, ApJL, 720, L82

    work page 2010

  8. [8]

    A., Liu, M

    Biller, B. A., Liu, M. C., Wahhaj, Z., et al. 2013, , 777, 160

    work page 2013

  9. [9]

    Binks , A. S. & Jeffries , R. D. 2017, , 469, 579

    work page 2017

  10. [10]

    2019, , 625, A21

    Boccaletti , A., Th \'e bault , P., Pawellek , N., et al. 2019, , 625, A21

    work page 2019

  11. [11]

    , Sezestre, E

    Boccaletti, A. , Sezestre, E. , Lagrange, A.-M. , et al. 2018, , 614, A52

    work page 2018

  12. [12]

    2022, , 663, A144

    Bonavita , M., Gratton , R., Desidera , S., et al. 2022, , 663, A144

    work page 2022

  13. [13]

    M., et al

    Bonnefoy , M., Perraut , K., Lagrange , A. M., et al. 2018, , 618, A63

    work page 2018

  14. [14]

    J., Gebhard, T

    Bonse, M. J., Gebhard, T. D., Dannert, F. A., et al. 2024, arXiv preprint arXiv:2406.01809

    work page arXiv 2024

  15. [15]

    2010, in Int

    Boureau, Y.-L., Ponce, J., & LeCun, Y. 2010, in Int. Conf. Mach. Learn. (ICML), Haifa, Israel, 111--118

    work page 2010

  16. [16]

    Bowler, B. P. 2016, Publications of the Astronomical Society of the Pacific, 128, 102001

    work page 2016

  17. [17]

    2020, in Adaptive Optics Systems VII, ed

    Cantalloube, F., Gomez-Gonzalez, C., Absil, O., et al. 2020, in Adaptive Optics Systems VII, ed. D. Schmidt, L. Schreiber, & E. Vernet (SPIE), 321

    work page 2020

  18. [18]

    2023, , 680, A86

    Cantero, C., Absil, O., Dahlqvist, C.-H., & Van Droogenbroeck, M. 2023, , 680, A86

    work page 2023

  19. [19]

    M., et al

    Chauvin , G., Desidera , S., Lagrange , A. M., et al. 2017, in SF2A-2017: Proceedings of the Annual meeting of the French Society of Astronomy and Astrophysics, ed. C. Reyl \'e , P. Di Matteo , F. Herpin , E. Lagadec , A. Lan c on , Z. Meliani , & F. Royer , Di

    work page 2017

  20. [20]

    2017, , 605, L1

    Chauvin, G., Desidera, S., Lagrange, A.-M., et al. 2017, , 605, L1

    work page 2017

  21. [21]

    2018, , 617, A76

    Chauvin , G., Gratton , R., Bonnefoy , M., et al. 2018, , 617, A76

    work page 2018

  22. [22]

    2004, , 425, L29–L32

    Chauvin, G., Lagrange, A.-M., Dumas, C., et al. 2004, , 425, L29–L32

    work page 2004

  23. [23]

    M., Lacombe , F., et al

    Chauvin , G., Lagrange , A. M., Lacombe , F., et al. 2005 a , , 430, 1027

    work page 2005

  24. [24]

    M., Zuckerman , B., et al

    Chauvin , G., Lagrange , A. M., Zuckerman , B., et al. 2005 b , , 438, L29

    work page 2005

  25. [25]

    , Vigan, A

    Chauvin, G. , Vigan, A. , Bonnefoy, M. , et al. 2015, , 573, A127

    work page 2015

  26. [26]

    2018, , 615, A160

    Cheetham , A., Bonnefoy , M., Desidera , S., et al. 2018, , 615, A160

    work page 2018

  27. [27]

    H., Mittal , T., Kuchner , M., et al

    Chen , C. H., Mittal , T., Kuchner , M., et al. 2014, , 211, 25

    work page 2014

  28. [28]

    2024, in Ground-based and Airborne Instrumentation for Astronomy X, ed

    Chilcote, J., Konopacky, Q., Hamper, R., et al. 2024, in Ground-based and Airborne Instrumentation for Astronomy X, ed. J. J. Bryant, K. Motohara, & J. R. D. Vernet, Vol. 13096, International Society for Optics and Photonics (SPIE), 1309699

    work page 2024

  29. [29]

    , Delorme, P

    Chomez, A. , Delorme, P. , Lagrange, A.-M. , et al. 2025, , 697, A99

    work page 2025

  30. [30]

    D., Chen, C

    Choquet, \'E ., Perrin, M. D., Chen, C. H., et al. 2016, ApJL, 817, L2

    work page 2016

  31. [31]

    Christiaens, V., Gonzalez, C. A. G., Farkas, R., et al. 2023, Journal of Open Source Software, 8, 4774

    work page 2023

  32. [32]

    U., Turatto , M., Gratton , R

    Claudi , R. U., Turatto , M., Gratton , R. G., et al. 2008, in Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series, Vol. 7014, Ground-based and Airborne Instrumentation for Astronomy II, ed. I. S. McLean & M. M. Casali , 70143E

    work page 2008

  33. [33]

    1987, , 181, 77

    Cote , J. 1987, , 181, 77

    work page 1987

  34. [34]

    P., Marcy, G

    Cumming, A., Butler, R. P., Marcy, G. W., et al. 2008, Publications of the Astronomical Society of the Pacific, 120, 531–554

    work page 2008

  35. [35]

    2024, , 692, A126

    Daglayan, H., Vary, S., Absil, O., et al. 2024, , 692, A126

    work page 2024

  36. [36]

    2022, in Int

    Daglayan, H., Vary, S., Cantalloube, F., Absil, P.-A., & Absil, O. 2022, in Int. Conf. Image Process. Appl. Syst. (IPAS), 1--5

    work page 2022

  37. [37]

    2020, , 633, 1

    Dahlqvist, C.-H., Cantalloube, F., & Absil, O. 2020, , 633, 1

    work page 2020

  38. [38]

    2021, , 646, 1

    Dahlqvist, C.-H., Louppe, G., & Absil, O. 2021, , 646, 1

    work page 2021

  39. [39]

    H., Milli , J., Absil , O., et al

    Dahlqvist , C. H., Milli , J., Absil , O., et al. 2022, , 666, A33

    work page 2022

  40. [40]

    2017, in Annual meeting of the French Society of Astronomy and Astrophysics (SF2A)

    Delorme , P., Meunier , N., Albert , D., et al. 2017, in Annual meeting of the French Society of Astronomy and Astrophysics (SF2A)

    work page 2017

  41. [41]

    2021, , 651, A70

    Desidera, S., Chauvin, G., Bonavita, M., et al. 2021, , 651, A70

    work page 2021

  42. [42]

    L., Meixner , M., Holfeltz , S

    Doering , R. L., Meixner , M., Holfeltz , S. T., et al. 2007, , 133, 2122

    work page 2007

  43. [43]

    2008, in Ground-based and Airborne Instrumentation for Astronomy II, Vol

    Dohlen , K., Langlois , M., Saisse , M., et al. 2008, in Ground-based and Airborne Instrumentation for Astronomy II, Vol. 7014, 70143L

    work page 2008

  44. [44]

    P., Mora , A., et al

    Eiroa , C., Marshall , J. P., Mora , A., et al. 2013, , 555, A11

    work page 2013

  45. [45]

    M., et al

    Engler , N., Boccaletti , A., Schmid , H. M., et al. 2019, , 622, A192

    work page 2019

  46. [46]

    2017, , 601, A7

    Feldt , M., Olofsson , J., Boccaletti , A., et al. 2017, , 601, A7

    work page 2017

  47. [47]

    2024, , 527, 1534

    Flasseur , O., Bodrito , T., Mairal , J., et al. 2024, , 527, 1534

    work page 2024

  48. [48]

    2011, , 526, A141

    Forveille , T., Bonfils , X., Lo Curto , G., et al. 2011, , 526, A141

    work page 2011

  49. [49]

    2006, Optics Express, 14, 7515

    Fusco, T., Rousset, G., Sauvage, J-.F annd Petit, C., et al. 2006, Optics Express, 14, 7515

    work page 2006

  50. [50]

    Gaia Collaboration , Vallenari , A., Brown , A. G. A., et al. 2023, , 674, A1

    work page 2023

  51. [51]

    2018, , 615, A92

    Galicher , R., Boccaletti , A., Mesa , D., et al. 2018, , 615, A92

    work page 2018

  52. [52]

    , Marois, C

    Galicher, R. , Marois, C. , Macintosh, B. , et al. 2016, , 594, A63

    work page 2016

  53. [53]

    2024, in Adaptive Optics Systems IX, ed

    Galland, N., Raynaud, H.-F., Kasper, M., et al. 2024, in Adaptive Optics Systems IX, ed. K. J. Jackson, D. Schmidt, & E. Vernet, Vol. 13097, International Society for Optics and Photonics (SPIE), 1309711

    work page 2024

  54. [54]

    2019, , 157, 39

    Gibbs, A., Wagner, K., Apai, D., et al. 2019, , 157, 39

    work page 2019

  55. [55]

    A., Tzanidakis, A., Wainer, T

    Gollotti, G. A., Tzanidakis, A., Wainer, T. M., & Davenport, J. R. A. 2024, Research Notes of the AAS, 8, 293

    work page 2024

  56. [56]

    2016, , 589, 1

    Gomez Gonzalez , C., Absil, O., Absil, P.-A., et al. 2016, , 589, 1

    work page 2016

  57. [57]

    2018, , 613, 1

    Gomez Gonzalez , C., Absil, O., & Van Droogenbroeck, M. 2018, , 613, 1

    work page 2018

  58. [58]

    2017, , 154, 7:1

    Gomez Gonzalez , C., Wertz, O., Absil, O., et al. 2017, , 154, 7:1

    work page 2017

  59. [59]

    , Caballero, J

    González-Payo, J. , Caballero, J. A. , Gorgas, J. , et al. 2024, , 689, A302

    work page 2024

  60. [60]

    & Migaszewski , C

    Go \'z dziewski , K. & Migaszewski , C. 2014, , 440, 3140

    work page 2014

  61. [61]

    , Bonavita, M

    Gratton, R. , Bonavita, M. , Mesa, D. , et al. 2024, , 685, A119

    work page 2024

  62. [62]

    Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. 2017, in Int. Conf. Mach. Learn. (ICML), Vol. 70, 1321–1330

    work page 2017

  63. [63]

    2021, in Int

    Gupta, K., Rahimi, A., Ajanthan, T., et al. 2021, in Int. Conf. Learn. Represent. (ICLR)

    work page 2021

  64. [64]

    D., et al

    Hendrycks, D., Mu, N., Cubuk, E. D., et al. 2020, in International Conference on Learning Representations

    work page 2020

  65. [65]

    & Belkin, M

    Hui, L. & Belkin, M. 2021, in Int. Conf. Learn. Represent. (ICLR)

    work page 2021

  66. [66]

    2021, , 646, A164

    Janson, M., Squicciarini, V., Delorme, P., et al. 2021, , 646, A164

    work page 2021

  67. [67]

    2021, A&A, 646, A164

    Janson , M., Squicciarini, Vito , Delorme, Philippe , et al. 2021, A&A, 646, A164

    work page 2021

  68. [68]

    2021, in Adv

    Karandikar, A., Cain, N., Tran, D., et al. 2021, in Adv. Neural Inf. Process. Syst. (NeurIPS), ed. A. Beygelzimer, Y. Dauphin, P. Liang, & J. W. Vaughan

    work page 2021

  69. [69]

    Kenworthy, M. A. & Haffert, S. Y. 2025, Annual Review of Astronomy and Astrophysics, 63, 179

    work page 2025

  70. [70]

    2018, , 617, 1

    Keppler, M., Benisty, M., M \"u ller, A., et al. 2018, , 617, 1

    work page 2018

  71. [71]

    M., Rameau , J., Duch \^e ne , G., et al

    Konopacky , Q. M., Rameau , J., Duch \^e ne , G., et al. 2016, , 829, L4

    work page 2016

  72. [72]

    2017, in Int

    Kull, M., Silva Filho, T., & Flach, P. 2017, in Int. Conf. Artif. Intell. Stat., 623--631

    work page 2017

  73. [73]

    2022, ApJL, 934, 1

    Kuzuhara, M., Currie, T., Takarada, T., et al. 2022, ApJL, 934, 1

    work page 2022

  74. [74]

    2011, , 730, 42

    Lafreni \`e re , D., Jayawardhana , R., Janson , M., et al. 2011, , 730, 42

    work page 2011

  75. [75]

    Lafreni \`e re , D., Jayawardhana, R., & van Kerkwijk, M. H. 2008, , 689, L153

    work page 2008

  76. [76]

    M., Gratadour, D., Chauvin, G., et al

    Lagrange, A. M., Gratadour, D., Chauvin, G., et al. 2009, , 493, L21

    work page 2009

  77. [77]

    M., Kiefer , F., Rubini , P., et al

    Lagrange , A. M., Kiefer , F., Rubini , P., et al. 2025 a , arXiv e-prints, arXiv:2501.10488

    work page arXiv 2025

  78. [78]

    M., Langlois , M., Gratton , R., et al

    Lagrange , A. M., Langlois , M., Gratton , R., et al. 2016, , 586, L8

    work page 2016

  79. [79]

    M., Meunier , N., Rubini , P., et al

    Lagrange , A. M., Meunier , N., Rubini , P., et al. 2019, Nature Astronomy, 3, 1135

    work page 2019

  80. [80]

    M., Wilkinson , C., M \^a lin , M., et al

    Lagrange , A. M., Wilkinson , C., M \^a lin , M., et al. 2025 b , , 642, 905

    work page 2025

Showing first 80 references.