pith. sign in

arxiv: 2606.07769 · v1 · pith:5XFT7MQAnew · submitted 2026-06-05 · 🌌 astro-ph.EP · astro-ph.IM· astro-ph.SR

Towards Instrument-Agnostic Exoplanet Candidate Prioritization

Vivaswan Kopparapu , Sibasish Laha , Brian P. Powell , Veselin Kostov , Rajarshi Basak , Samuel Verbrugge This is my paper

Pith reviewed 2026-06-27 20:42 UTC · model grok-4.3

classification 🌌 astro-ph.EP astro-ph.IMastro-ph.SR
keywords exoplanet candidate prioritizationmachine learningTESSKeplertransit parametersensemble modelinstrument-agnostic
0
0 comments X

The pith

Joint training on Kepler and TESS data allows machine learning models to prioritize exoplanet candidates effectively on both instruments.

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

The paper shows that standard machine learning models struggle when trained on data from only one telescope because the distributions of key parameters differ between Kepler and TESS. When the same models are instead trained on the combined datasets, they achieve good performance on candidates from either instrument. The authors select six transit and stellar parameters, test eleven classifiers across all train-test splits, and assemble the best performers into an ensemble that ranks remaining candidates. They release the top-ranked lists and note that newly confirmed planets since their analysis align with the model's predictions. The work positions this joint-training approach as a practical way to handle the much larger candidate volume expected from the Roman Space Telescope.

Core claim

Models trained jointly with both TESS and Kepler data can perform well on both, whereas models trained on data from only one instrument have difficulty predicting the other because of substantially different distributions in the chosen parameters; an ensemble of the best such jointly trained models can therefore be used to rank planet candidates in either archive.

What carries the argument

Ensemble of eleven machine-learning classifiers trained on the six parameters (orbital period, planet radius, stellar temperature, stellar radius, transit depth, transit duration) using combined Kepler and TESS labels.

If this is right

  • Joint training removes the need to build separate instrument-specific models for candidate vetting.
  • The same six-parameter feature set and ensemble can be applied directly to new candidates from either archive.
  • Top-ranked candidates from the model are more likely to be confirmed upon further observation.
  • The approach scales to the order-of-magnitude increase in candidates expected from the Roman Space Telescope.

Where Pith is reading between the lines

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

  • The method could be tested on simulated Roman data to check whether the same six parameters remain sufficient when the instrument changes again.
  • If the joint-training benefit holds, similar cross-mission ensembles might reduce the need for mission-specific training sets in other time-domain surveys.
  • Discrepancies between the model's top candidates and follow-up results would point to missing parameters or label noise rather than instrument-specific effects.

Load-bearing premise

The archival labels of confirmed planets versus false positives are accurate enough to serve as reliable training targets, and the six chosen parameters capture the essential distinguishing features across instruments.

What would settle it

A list of newly confirmed planets or false positives from either Kepler or TESS that were not used in training; if the ensemble's rankings systematically disagree with these new labels on one instrument but not the other, the joint-training claim fails.

Figures

Figures reproduced from arXiv: 2606.07769 by Brian P. Powell, Rajarshi Basak, Samuel Verbrugge, Sibasish Laha, Veselin Kostov, Vivaswan Kopparapu.

Figure 1
Figure 1. Figure 1: Schematic showing a planet transiting a star. The six parameters that we used from the TESS and Kepler databases to train our ML models are noted in the diagram. NExScI can be found on their separate websites for TESS4 and Kepler5 . For the TESS dataset, we selected the TESS TOI catalog (current as of August 30, 2025) that includes false positives (FP), planet candidates (PC), ambiguous planetary candidate… view at source ↗
Figure 2
Figure 2. Figure 2: Distribution of input features separated by source (TESS/Kepler) and label (confirmed planets/false positives). TESS CP+KP (n=1198, dark blue), TESS FA+FP (n=1137, light blue), Kepler CP (n=2744, dark orange), Kepler FP (n=4582, light orange). All features except Teff are plotted on a logarithmic x-axis. KS test results between the full distributions (i.e. all TESS and all Kepler) are noted in a box below … view at source ↗
Figure 3
Figure 3. Figure 3: Heatmaps showing mean classifier performance across all train/test configurations using 20-iteration repeated random sub-sampling. Color scale ranges from 0.0 (red) to 1.0 (green). Panels show (top-left to bottom-right): Accuracy, Precision, Recall, F1, PR-AUC, and ROC-AUC. For scale-sensitive models, subscripts ‘u’ and ‘s’ indicate unscaled and scaled, respectively. Train/test configuration abbreviations … view at source ↗
Figure 4
Figure 4. Figure 4: Distribution of ensemble model prediction scores for planet candidates from TESS and Kepler. (left) TESS candidates (n=4701). (center) Kepler candidates (n=1874). (right) Normalized overlay comparison. Each candidate’s overall score is the mean of 60 predictions (3 models × 20 iterations) from MLP, RF, and XGB classifiers trained on balanced confirmed planets and false positives from both surveys. ensemble… view at source ↗
Figure 5
Figure 5. Figure 5: Pairwise comparison of model predictions for TESS (blue) and Kepler (orange) planet candidates. (left) RF vs XGB, (middle) RF vs MLP, and (right) XGB vs MLP. Each panel shows the predicted probability from one model versus another. Dashed diagonal line indicates perfect agreement. Correlation coefficients for TESS and Kepler are given by r(T) and r(K) on each plot, respectively. 0.0 0.2 0.4 0.6 0.8 1.0 Ove… view at source ↗
Figure 6
Figure 6. Figure 6: Ensemble model scores for 203 TOIs resolved as either CP (green) or FP (red) between the dates of the data we used for our analysis (August 30, 2025 and April 28, 2026). See Section 4.5 for details [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
read the original abstract

We have developed a novel machine learning (ML) approach for predicting the likelihood of exoplanet candidate confirmation equally capable of performance on both TESS and Kepler data. From the NASA exoplanet archival post-processed Kepler and TESS databases, we chose six parameters that we assessed to be predictive to the planet transit signature: planet orbital period (P), planet radius ($R_{\rm p}$), stellar temperature ($T_{\rm eff}$), stellar radius ($R_{\rm star}$), planet transit depth ($\delta$), and planet transit duration ($t_{\rm d}$). We used these parameters to evaluate eleven different ML models on all possible train/test combinations of TESS and Kepler data, using the confirmed planet and false positive labels as our training targets. We found that, due to substantially different distributions of our chosen parameters in Kepler and TESS databases, models trained with data from one instrument have difficulty predicting the other. However, models trained jointly with both TESS and Kepler data can perform well on both. We combined our best models into a statistically robust ensemble to evaluate the planet candidates in both Kepler and TESS, and we provide a list of the top candidates predicted by our model for each. Confirmed planets and false positives that have been resolved since the completion of our analysis demonstrate the effectiveness of our model and suggest that our top candidates are likely to be confirmed if they are further analyzed by the community. With the upcoming launch of the Nancy Grace Roman Space Telescope (Roman) and the expected order-of-magnitude increase in planet candidates, we suggest that our method can be extended to Roman data for robust and effective prioritization for analysis.

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 / 2 minor

Summary. The manuscript presents an empirical ML study that selects six parameters (orbital period P, planet radius Rp, stellar effective temperature Teff, stellar radius Rstar, transit depth δ, transit duration td) from NASA exoplanet archives and trains eleven models on all combinations of Kepler and TESS data, using archival confirmed-planet versus false-positive labels as targets. It reports that single-instrument training fails to generalize across the differing parameter distributions, while joint training succeeds on both instruments; an ensemble of the best models is then applied to rank candidates in each survey, with a short list of top candidates provided and anecdotal post-analysis confirmations cited as supporting evidence. The work suggests the method can be extended to future missions such as Roman.

Significance. If the joint-training improvement can be shown to be robust to label noise and distribution shifts, the approach offers a practical, instrument-agnostic prioritization tool that could scale to the order-of-magnitude increase in candidates expected from Roman. The use of standard archival labels and off-the-shelf models makes the method immediately reproducible, but the absence of reported performance numbers means the claimed advantage remains unevaluated.

major comments (3)
  1. [Abstract] Abstract and the paragraph reporting results: the central claim that 'models trained jointly with both TESS and Kepler data can perform well on both' is stated without any accompanying quantitative metrics (accuracy, precision, recall, F1, AUC, or cross-validation scores), error bars, or ablation tables. Without these numbers the empirical result cannot be assessed and the comparison to single-instrument training remains qualitative.
  2. [Methods / data preparation] Section describing label usage and training targets: the manuscript relies directly on archival confirmed-planet versus false-positive labels without any audit of label noise, completeness, or instrument-specific systematics (e.g., Kepler’s longer baseline and more extensive follow-up versus TESS). If label errors correlate with the six chosen parameters or with instrument, the apparent benefit of joint training could be an artifact rather than a reflection of intrinsic transit signatures.
  3. [Results] Results section on ensemble ranking: no feature-importance analysis, ablation on the six parameters, or comparison against simple threshold baselines on the same parameters is reported. This leaves open whether the ML ensemble adds predictive power beyond the input features themselves.
minor comments (2)
  1. [Introduction / parameter selection] The six parameters are introduced with the phrase 'we assessed to be predictive' but no supporting correlation table or justification appears in the text.
  2. [Abstract] Notation for transit depth (δ) and duration (td) should be defined at first use and kept consistent with standard exoplanet literature.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive comments, which highlight areas where the manuscript can be strengthened with additional quantitative details and analyses. We address each major comment below and will revise the manuscript accordingly to include the requested metrics, discussions, and comparisons.

read point-by-point responses

  1. Referee: [Abstract] Abstract and the paragraph reporting results: the central claim that 'models trained jointly with both TESS and Kepler data can perform well on both' is stated without any accompanying quantitative metrics (accuracy, precision, recall, F1, AUC, or cross-validation scores), error bars, or ablation tables. Without these numbers the empirical result cannot be assessed and the comparison to single-instrument training remains qualitative.

    Authors: We agree that quantitative metrics are necessary to substantiate the central claim. In the revised manuscript we will report accuracy, precision, recall, F1, and AUC scores (with standard deviations from 5-fold cross-validation) for all eleven models across the four train/test combinations (Kepler-only, TESS-only, joint training on both, and joint testing on each). These numbers will be added to both the abstract and the results section, together with a compact table comparing single-instrument versus joint-training performance. revision: yes

  2. Referee: [Methods / data preparation] Section describing label usage and training targets: the manuscript relies directly on archival confirmed-planet versus false-positive labels without any audit of label noise, completeness, or instrument-specific systematics (e.g., Kepler’s longer baseline and more extensive follow-up versus TESS). If label errors correlate with the six chosen parameters or with instrument, the apparent benefit of joint training could be an artifact rather than a reflection of intrinsic transit signatures.

    Authors: We acknowledge that archival labels carry potential noise and instrument-dependent biases. A full independent audit of label completeness and systematics lies outside the scope of the present empirical study, which deliberately uses the standard NASA Exoplanet Archive labels to ensure immediate reproducibility. In revision we will add an explicit limitations paragraph discussing possible label noise, its possible correlation with the chosen parameters, and the fact that any such noise would affect both single- and joint-training regimes. We will also note that the observed cross-instrument generalization remains an empirical result even if label quality varies. revision: partial

  3. Referee: [Results] Results section on ensemble ranking: no feature-importance analysis, ablation on the six parameters, or comparison against simple threshold baselines on the same parameters is reported. This leaves open whether the ML ensemble adds predictive power beyond the input features themselves.

    Authors: We agree that these analyses would strengthen the claim that the ensemble contributes beyond the raw parameters. The revised manuscript will include (i) permutation-based feature-importance rankings averaged across the best models, (ii) an ablation table showing performance when each of the six parameters is removed in turn, and (iii) direct comparisons of the ensemble against simple threshold baselines (e.g., cuts on transit depth, period, and radius) applied to the same test sets. These additions will quantify the incremental value of the learned models. revision: yes

Circularity Check

0 steps flagged

No circularity; empirical ML on external archival labels

full rationale

The paper trains standard ML classifiers on six transit-related parameters using NASA archival confirmed-planet versus false-positive labels as targets. All reported results are obtained from explicit train/test splits (including cross-instrument combinations) and evaluated on held-out data plus post-analysis resolved cases. No derivations, equations, uniqueness theorems, or self-citations are invoked to justify the central claims; the pipeline is self-contained against external benchmarks with no reduction of predictions to fitted inputs by construction.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on the domain assumption that the six transit parameters are sufficient to distinguish planets from false positives and that archival labels are clean enough for supervised learning; no new physical entities are introduced and model hyperparameters are implicit free parameters not enumerated in the abstract.

free parameters (1)
  • ML model hyperparameters for the eleven models
    Choice and tuning of the eleven different ML models are required to achieve the reported joint-training performance.
axioms (2)
  • domain assumption The six selected parameters capture the essential differences between confirmed planets and false positives
    Paper states these parameters were assessed to be predictive.
  • domain assumption Archival confirmed-planet and false-positive labels constitute reliable supervised training targets
    Labels are used directly as training targets without additional validation described.

pith-pipeline@v0.9.1-grok · 5856 in / 1310 out tokens · 38334 ms · 2026-06-27T20:42:28.826351+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

76 extracted references · 43 canonical work pages · 18 internal anchors

  1. [1]

    Identifying Exoplanets with Deep Learning: A Five Planet Resonant Chain around Kepler-80 and an Eighth Planet around Kepler-90

    Identifying Exoplanets with Deep Learning: A Five-planet Resonant Chain around Kepler-80 and an Eighth Planet around Kepler-90. , keywords =. doi:10.3847/1538-3881/aa9e09 , archivePrefix =. 1712.05044 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-3881/aa9e09

  2. [2]

    Science , volume=

    Kepler planet-detection mission: introduction and first results , author=. Science , volume=. 2010 , publisher=

    2010

  3. [3]

    Journal of Machine Learning Research , volume=

    Scikit-learn: Machine Learning in Python , author=. Journal of Machine Learning Research , volume=. 2011 , url=

    2011

  4. [4]

    The Astrophysical Journal Letters , volume=

    Kepler mission design, realized photometric performance, and early science , author=. The Astrophysical Journal Letters , volume=. 2010 , publisher=

    2010

  5. [5]

    The Astrophysical Journal Letters , volume=

    Initial characteristics of Kepler long cadence data for detecting transiting planets , author=. The Astrophysical Journal Letters , volume=. 2010 , publisher=

    2010

  6. [6]

    Caldwell, D. A. and et al. , title =. ApJL , year =

  7. [7]

    Ricker, G. R. and et al. , title =. JATIS , year =

  8. [8]

    , title =

    Massey, Frank J. , title =. Journal of the American Statistical Association , year =

  9. [9]

    Identifying Exoplanets with Deep Learning. III. Automated Triage and Vetting of TESS Candidates. , keywords =. doi:10.3847/1538-3881/ab21d6 , archivePrefix =. 1904.02726 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-3881/ab21d6 1904

  10. [10]

    , keywords =

    ExoMiner: A Highly Accurate and Explainable Deep Learning Classifier That Validates 301 New Exoplanets. , keywords =. doi:10.3847/1538-4357/ac4399 , archivePrefix =. 2111.10009 , primaryClass =

    work page doi:10.3847/1538-4357/ac4399

  11. [11]

    , keywords =

    Identification of the Top TESS Objects of Interest for Atmospheric Characterization of Transiting Exoplanets with JWST. , keywords =. doi:10.3847/1538-3881/ad3068 , archivePrefix =. 2308.09617 , primaryClass =

    work page doi:10.3847/1538-3881/ad3068

  12. [12]

    Identifying Exoplanets with Deep Learning. V. Improved Light-curve Classification for TESS Full-frame Image Observations. , keywords =. doi:10.3847/1538-3881/acad85 , archivePrefix =. 2301.01371 , primaryClass =

    work page doi:10.3847/1538-3881/acad85

  13. [13]

    Experimental Astronomy , keywords =

    The PLATO mission. Experimental Astronomy , keywords =. doi:10.1007/s10686-025-09985-9 , archivePrefix =. 2406.05447 , primaryClass =

    work page doi:10.1007/s10686-025-09985-9

  14. [14]

    , keywords =

    Exoplanet validation with machine learning: 50 new validated Kepler planets. , keywords =. doi:10.1093/mnras/staa2498 , archivePrefix =. 2008.10516 , primaryClass =

    work page doi:10.1093/mnras/staa2498 2008

  15. [15]

    Predictions of the WFIRST Microlensing Survey. I. Bound Planet Detection Rates. , keywords =. doi:10.3847/1538-4365/aafb69 , archivePrefix =. 1808.02490 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-4365/aafb69

  16. [16]

    , keywords =

    ExoMiner++: Enhanced Transit Classification and a New Vetting Catalog for 2-minute TESS Data. , keywords =. doi:10.3847/1538-3881/ae03a4 , archivePrefix =. 2502.09790 , primaryClass =

    work page doi:10.3847/1538-3881/ae03a4

  17. [17]

    , keywords =

    Multiplicity Boost of Transit Signal Classifiers: Validation of 69 New Exoplanets using the Multiplicity Boost of ExoMiner. , keywords =. doi:10.3847/1538-3881/acd344 , archivePrefix =. 2305.02470 , primaryClass =

    work page doi:10.3847/1538-3881/acd344

  18. [18]

    Validation of Kepler's Multiple Planet Candidates. II. Refined Statistical Framework and Descriptions of Systems of Special Interest. , keywords =. doi:10.1088/0004-637X/784/1/44 , archivePrefix =. 1402.6352 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/784/1/44

  19. [19]

    Almost All of Kepler's Multiple Planet Candidates are Planets

    Almost All of Kepler's Multiple-planet Candidates Are Planets. , keywords =. doi:10.1088/0004-637X/750/2/112 , archivePrefix =. 1201.5424 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/750/2/112

  20. [20]

    arXiv e-prints , keywords =

    WATSON-Net: Vetting, Validation, and Analysis of Transits from Space Observations with Neural Networks. arXiv e-prints , keywords =. doi:10.48550/arXiv.2511.08768 , archivePrefix =. 2511.08768 , primaryClass =

    work page doi:10.48550/arxiv.2511.08768

  21. [21]

    , keywords =

    Characterization of two new transiting sub-Neptunes and a terrestrial planet around M-dwarf hosts. , keywords =. doi:10.1051/0004-6361/202557155 , archivePrefix =. 2601.07414 , primaryClass =

    work page doi:10.1051/0004-6361/202557155

  22. [22]

    Automatic search for transiting planets in TESS-SPOC FFIs with RAVEN: over 100 newly validated planets and over 2000 vetted candidates

    Automatic search for transiting planets in TESS-SPOC FFIs with RAVEN: over 100 newly validated planets and over 2000 vetted candidates. , keywords =. doi:10.1093/mnras/stag512 , archivePrefix =. 2603.22597 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stag512 2000

  23. [23]

    arXiv e-prints , keywords =

    RAVEN: RAnking and Validation of ExoplaNets. arXiv e-prints , keywords =. doi:10.48550/arXiv.2509.17645 , archivePrefix =. 2509.17645 , primaryClass =

    work page doi:10.48550/arxiv.2509.17645

  24. [24]

    Discovery and Vetting of Exoplanets. I. Benchmarking K2 Vetting Tools. , keywords =. doi:10.3847/1538-3881/ab0110 , archivePrefix =. 1901.07459 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-3881/ab0110 1901

  25. [25]

    Planetary Candidates Observed by Kepler. VIII. A Fully Automated Catalog with Measured Completeness and Reliability Based on Data Release 25. , keywords =. doi:10.3847/1538-4365/aab4f9 , archivePrefix =. 1710.06758 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-4365/aab4f9

  26. [26]

    The Astronomical Journal , volume =

    Kunimoto, Michelle and others , title =. The Astronomical Journal , volume =. 2025 , month =

    2025

  27. [27]

    A Machine Learning Technique to Identify Transit Shaped Signals

    A Machine Learning Technique to Identify Transit Shaped Signals. , keywords =. doi:10.1088/0004-637X/812/1/46 , archivePrefix =. 1509.00041 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/812/1/46

  28. [28]

    Automatic Classification of Kepler Planetary Transit Candidates

    Automatic Classification of Kepler Planetary Transit Candidates. , keywords =. doi:10.1088/0004-637X/806/1/6 , archivePrefix =. 1408.1496 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/806/1/6

  29. [29]

    , keywords =

    Vetting of 384 TESS Objects of Interest with TRICERATOPS and Statistical Validation of 12 Planet Candidates. , keywords =. doi:10.3847/1538-3881/abc6af , archivePrefix =. 2002.00691 , primaryClass =

    work page doi:10.3847/1538-3881/abc6af 2002

  30. [30]

    An Efficient Automated Validation Procedure for Exoplanet Transit Candidates

    An Efficient Automated Validation Procedure for Exoplanet Transit Candidates. , keywords =. doi:10.1088/0004-637X/761/1/6 , archivePrefix =. 1206.1568 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/761/1/6

  31. [31]

    , keywords =

    An ultra-short period super-Earth and sub-Neptune spanning the Radius Valley orbiting the kinematic thick disc star TOI-2345. , keywords =. doi:10.1093/mnras/staf1806 , archivePrefix =. 2510.12783 , primaryClass =

    work page doi:10.1093/mnras/staf1806

  32. [32]

    TESS Observatory Guide v1.1 , howpublished =

  33. [33]

    TESS Science: Observations (cadences and FFIs) , howpublished =

  34. [34]

    Jenkins, J. M. and et al. , title =. SPIE Software and Cyberinfrastructure for Astronomy IV , year =

  35. [35]

    Huang, C. X. and et al. , title =. RNAAS , year =

  36. [36]

    and et al

    Kunimoto, M. and et al. , title =. arXiv , year =

  37. [37]

    and et al

    Kunimoto, M. and et al. , title =. RNAAS , year =

  38. [38]

    Burke, C. J. and et al. , title =. ApJ , year =

  39. [39]

    Hsu, D. C. and Ford, E. B. and Ragozzine, D. and Morehead, R. C. , title =. AJ , year =

  40. [40]

    and et al

    Bryson, S. and et al. , title =. AJ , year =

  41. [41]

    and Matthews, J

    Kunimoto, M. and Matthews, J. M. , title =. AJ , year =

  42. [42]

    , keywords =

    The Occurrence Rate of Terrestrial Planets Orbiting Nearby Mid-to-late M Dwarfs from TESS Sectors 1-42. , keywords =. doi:10.3847/1538-3881/acd175 , archivePrefix =. 2302.04242 , primaryClass =

    work page doi:10.3847/1538-3881/acd175

  43. [43]

    The Gaia-Kepler Stellar Properties Catalog. II. Planet Radius Demographics as a Function of Stellar Mass and Age. , keywords =. doi:10.3847/1538-3881/aba18a , archivePrefix =. 2005.14671 , primaryClass =

    work page doi:10.3847/1538-3881/aba18a 2005

  44. [44]

    , keywords =

    The TESS Objects of Interest Catalog from the TESS Prime Mission. , keywords =. doi:10.3847/1538-4365/abefe1 , archivePrefix =. 2103.12538 , primaryClass =

    work page doi:10.3847/1538-4365/abefe1

  45. [45]

    Bryant, E. M. and Bayliss, D. and Van Eylen, V. , title =. MNRAS , year =

  46. [46]

    Howell, S. B. and et al. , title =. PASP , year =

  47. [47]

    Henderson, C. B. and et al. , title =. PASP , year =

  48. [48]

    and Johnson, J

    Vanderburg, A. and Johnson, J. A. , title =. PASP , year =

  49. [49]

    and et al

    Luger, R. and et al. , title =. AJ , year =

  50. [50]

    Crossfield, I. J. M. and et al. , title =. ApJ , year =

  51. [51]

    Montet, B. T. and et al. , title =. ApJ , year =

  52. [52]

    The Astronomical Journal , volume=

    No Evidence for More Earth-sized Planets in the Habitable Zone of Kepler's M versus FGK Stars , author=. The Astronomical Journal , volume=. 2023 , publisher=

    2023

  53. [53]

    The Astronomical Journal , volume=

    The occurrence rate of terrestrial planets orbiting nearby mid-to-late M dwarfs from TESS sectors 1--42 , author=. The Astronomical Journal , volume=. 2023 , publisher=

    2023

  54. [54]

    CASA, the Common Astronomy Software Applications for Radio Astronomy

    CASA, the Common Astronomy Software Applications for Radio Astronomy. , keywords =. doi:10.1088/1538-3873/ac9642 , archivePrefix =. 2210.02276 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/1538-3873/ac9642

  55. [55]

    Indecent Exposure in Seyfert 2 Galaxies: A Close Look

    Indecent Exposure in Seyfert 2 Galaxies: A Close Look. , keywords =. doi:10.1088/2041-8205/726/2/L21 , archivePrefix =. 1012.1865 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/2041-8205/726/2/l21 2041

  56. [56]

    , keywords =

    A Reemerging Bright Soft X-Ray State of the Changing-look Active Galactic Nucleus 1ES 1927+654: A Multiwavelength View. , keywords =. doi:10.3847/1538-4357/aced92 , archivePrefix =. 2308.03602 , primaryClass =

    work page doi:10.3847/1538-4357/aced92 1927

  57. [57]

    , keywords =

    Rapid luminosity decline and subsequent reformation of the innermost dust distribution in the changing-look AGN Mrk 590. , keywords =. doi:10.1093/mnras/stz3397 , archivePrefix =. 1904.08946 , primaryClass =

    work page doi:10.1093/mnras/stz3397 1904

  58. [58]

    , keywords =

    Lifting the curtain: The Seyfert galaxy Mrk 335 emerges from deep low-state in a sequence of rapid flare events. , keywords =. doi:10.1051/0004-6361/202039098 , archivePrefix =. 2011.04996 , primaryClass =

    work page doi:10.1051/0004-6361/202039098 2011

  59. [59]

    , keywords =

    A Radio, Optical, UV, and X-Ray View of the Enigmatic Changing-look Active Galactic Nucleus 1ES 1927+654 from Its Pre- to Postflare States. , keywords =. doi:10.3847/1538-4357/ac63aa , archivePrefix =. 2203.07446 , primaryClass =

    work page doi:10.3847/1538-4357/ac63aa 1927

  60. [60]

    The Changing-Look Quasar Mrk 590 is Awakening

    The Changing-look Quasar Mrk 590 Is Awakening. , keywords =. doi:10.3847/1538-4357/aadd91 , archivePrefix =. 1810.06616 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-4357/aadd91

  61. [61]

    Astronomical Data Analysis Software and Systems XVI , year = 2007, editor =

    CASA Architecture and Applications. Astronomical Data Analysis Software and Systems XVI , year = 2007, editor =

    2007

  62. [62]

    15th European VLBI Network Mini-Symposium and Users' Meeting , year = 2023, month = aug, eid =

    ngDIFMAP : new generation DIFMAP for Modelfitting Interferometric Data. 15th European VLBI Network Mini-Symposium and Users' Meeting , year = 2023, month = aug, eid =

    2023

  63. [63]

    , keywords =

    1ES 1927+654: An AGN Caught Changing Look on a Timescale of Months. , keywords =. doi:10.3847/1538-4357/ab39e4 , archivePrefix =. 1903.11084 , primaryClass =

    work page doi:10.3847/1538-4357/ab39e4 1927

  64. [64]

    SOAR TESS Survey. II. The Impact of Stellar Companions on Planetary Populations. , keywords =. doi:10.3847/1538-3881/ac17f6 , archivePrefix =. 2103.12076 , primaryClass =

    work page doi:10.3847/1538-3881/ac17f6

  65. [65]

    The California-Kepler Survey. III. A Gap in the Radius Distribution of Small Planets. , keywords =. doi:10.3847/1538-3881/aa80eb , archivePrefix =. 1703.10375 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/1538-3881/aa80eb

  66. [66]

    The false positive rate of Kepler and the occurrence of planets

    The False Positive Rate of Kepler and the Occurrence of Planets. , keywords =. doi:10.1088/0004-637X/766/2/81 , archivePrefix =. 1301.0842 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/766/2/81

  67. [67]

    The Transiting Exoplanet Survey Satellite: Simulations of planet detections and astrophysical false positives

    The Transiting Exoplanet Survey Satellite: Simulations of Planet Detections and Astrophysical False Positives. , keywords =. doi:10.1088/0004-637X/809/1/77 , archivePrefix =. 1506.03845 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/809/1/77

  68. [68]

    Proceedings of the 14th International Joint Conference on Artificial Intelligence (IJCAI) , year =

    Kohavi, Ron , title =. Proceedings of the 14th International Joint Conference on Artificial Intelligence (IJCAI) , year =

  69. [69]

    , keywords =

    Transiting Exoplanet Yields for the Roman Galactic Bulge Time Domain Survey Predicted from Pixel-level Simulations. , keywords =. doi:10.3847/1538-4365/acf3df , archivePrefix =. 2305.16204 , primaryClass =

    work page doi:10.3847/1538-4365/acf3df

  70. [70]

    arXiv e-prints , keywords =

    Sibling Sub-Neptunes Around Sibling M Dwarfs: TOI-521 and TOI-912. arXiv e-prints , keywords =. doi:10.48550/arXiv.2509.14782 , archivePrefix =. 2509.14782 , primaryClass =

    work page doi:10.48550/arxiv.2509.14782

  71. [71]

    OrCAS: Origins, Compositions, and Atmospheres of Sub-Neptunes. I. Survey Definition. , keywords =. doi:10.3847/1538-3881/ad9aa6 , archivePrefix =. 2411.16836 , primaryClass =

    work page doi:10.3847/1538-3881/ad9aa6

  72. [72]

    High five from ASTEP: Three validated planets and two eclipsing binaries in a diverse set of long-period candidates

    High Five From ASTEP: Three Validated Planets and Two Eclipsing Binaries in a Diverse Set of Long-Period Candidates. arXiv e-prints , keywords =. doi:10.48550/arXiv.2510.01725 , archivePrefix =. 2510.01725 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2510.01725

  73. [73]

    , keywords =

    From Earths to Super-Earths: Five New Small Planets Transiting M Dwarf Stars. , keywords =. doi:10.3847/1538-3881/ae246f , archivePrefix =. 2512.11971 , primaryClass =

    work page doi:10.3847/1538-3881/ae246f

  74. [74]

    Advances in Neural Information Processing Systems 30 , series =

    Ke, Guolin and Meng, Qi and Finley, Thomas and Wang, Taifeng and Chen, Wei and Ma, Weidong and Ye, Qiwei and Liu, Tie-Yan , title =. Advances in Neural Information Processing Systems 30 , series =. 2017 , pages =

    2017

  75. [75]

    Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , series =

    Chen, Tianqi and Guestrin, Carlos , title =. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , series =. 2016 , pages =

    2016

  76. [76]

    Astronomical Data Analysis Software and Systems V , year = 1996, editor =

    AIPS Developments in the Nineties. Astronomical Data Analysis Software and Systems V , year = 1996, editor =

    1996