pith. sign in

arxiv: 2606.11737 · v1 · pith:MLKZTV2Znew · submitted 2026-06-10 · 🌌 astro-ph.EP · astro-ph.IM· cs.LG

Machine-learning clustering of close-in exoplanet populations: links to pebble accretion

Yi Duann , Anders Johansen , Haiyang S. Wang , H. Jens Hoeijmakers This is my paper

Pith reviewed 2026-06-27 08:35 UTC · model grok-4.3

classification 🌌 astro-ph.EP astro-ph.IMcs.LG
keywords exoplanetspebble accretionmachine learningclusteringplanet formationgas giantsclose-in planets
0
0 comments X

The pith

Clustering reveals early formation for massive close-in giants

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

This paper applies a two-stage Gaussian mixture model to dynamical parameters of observed close-in exoplanets to find natural groupings. These groups are then mapped to synthetic populations from pebble accretion models. The mapping shows systematic differences in formation timing, with very-massive gas giants forming earlier than hot giants and warm Jupiters. The approach avoids predefined categories and uses physical parameters to link data to theory. Readers may find it useful for understanding how formation conditions shape the exoplanets we observe today.

Core claim

The two-stage GMM clustering in a feature space of dynamical descriptors identifies sub-populations of close-in exoplanets. When mapped to pebble-accretion synthetic populations in a three-dimensional parameter space, these reveal differences in formation timing, gas accretion, and solid growth histories, with very-massive gas giants preferentially linked to earlier formation epochs than the hot-giant and warm-Jupiter populations.

What carries the argument

Two-stage Gaussian mixture model for probabilistic clustering of dynamical planet-star interaction descriptors, mapped onto pebble-accretion formation models via gas availability, gas fraction, and ice-rock mass ratio.

If this is right

  • Very-massive gas giants form at earlier epochs than hot giants and warm Jupiters.
  • Clusters show distinct gas accretion and solid growth histories.
  • The mapping provides a statistical link between observed populations and theoretical formation pathways.
  • Sub-populations emerge without imposed classification boundaries.

Where Pith is reading between the lines

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

  • This clustering could be extended to include more exoplanet properties to test robustness against migration models.
  • Independent age measurements of exoplanet systems might confirm the inferred formation epoch differences.
  • The framework suggests that formation timing is a key driver of observed close-in architectures.

Load-bearing premise

The clusters from dynamical parameters reflect distinct pebble-accretion pathways rather than being primarily influenced by observational biases or post-formation migration.

What would settle it

If incorporating migration into the synthetic population models causes the mapped clusters to lose their distinct formation timing signatures, or if direct age determinations contradict the earlier formation for very-massive giants.

Figures

Figures reproduced from arXiv: 2606.11737 by Anders Johansen, Haiyang S. Wang, H. Jens Hoeijmakers, Yi Duann.

Figure 1
Figure 1. Figure 1: Distributions of the nine input parameters before (blue) and after (orange) Z-score filtering, shown in separate panels for each parameter. The bottom axis in each panel indicates the standardised value used in the GMM analysis, while the top axis shows the corresponding physical value in the original units (see [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Schematic overview of the analysis pipeline adopted in this work, in which planets from the NASA Exoplanet Archive are clus￾tered using machine-learning techniques and subsequently mapped into a three-dimensional parameter space using a simulation-based synthetic population. Parameters shown in blue denote quantities that are not di￾rectly accessible from current observations and are therefore defined only… view at source ↗
Figure 3
Figure 3. Figure 3: b. The maximum posterior cluster membership prob￾ability exceeds 0.8 for most objects, indicating that individual planets are typically well associated with a dominant mixture component in a given realisation of the GMM. In contrast, the cluster assignment stability estimated via bootstrap resampling is typically distributed in the range 0.2–0.4, reflecting that a sub￾set of objects may switch between neig… view at source ↗
Figure 5
Figure 5. Figure 5: focuses on the second-stage subdivision of cluster A into sub-clusters A1 and A2. The normalised profiles of these sub-clusters reveal differences within the originally dominant population, while the associated category distributions show how the second-stage GMM further partitions planets that were grouped together in the first-stage classification. A comparison between Figures 4 and 5 highlights the dist… view at source ↗
Figure 4
Figure 4. Figure 4: Normalised parameter profiles for the four clusters (A–D) iden￾tified in the first-stage GMM classification. In each panel, grey curves show the normalised parameter vectors of individual planets assigned to the corresponding cluster, while the black solid line indicates the me￾dian profile of the cluster. The percentage quoted in each panel title denotes the fraction of the full sample assigned to that cl… view at source ↗
Figure 6
Figure 6. Figure 6: Relative contributions of the final GMM clusters (A1, A2, B, C, and D) within [PITH_FULL_IMAGE:figures/full_fig_p007_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Logarithmic plane of ρp–qp/s (Top) and a–qp/s (Bottom). Filled circles denote planets from the NASA Exoplanet Archive, coloured by their GMM cluster assignments (A1, A2, B, C, and D), with error bars indicating measurement uncertainties. Planets with low clustering con￾fidence are marked by triangles. ciated with large mass ratios, while cluster B occupies a regime of lower masses and higher densities. Clu… view at source ↗
Figure 7
Figure 7. Figure 7: Normalised parameter profiles for six planets classified as low￾confidence cases in the GMM analysis. For each object, the red curve shows the normalised planetary parameter vector, while the black solid line indicates the median profile of the primary GMM cluster. The blue dashed line corresponds to the median profile of the second most prob￾able cluster based on posterior membership probabilities. 4.3. P… view at source ↗
Figure 9
Figure 9. Figure 9: Distribution of synthetic planets in the log10(qp/s)–Ice–Rock ra￾tio space. Each point represents a synthetic planet, color-coded by its formation time tt0 in Myr. The inset panel shows the normalised his￾togram of Ice–Rock ratios for the entire population, with the median value indicated by the red vertical line. The distribution is strongly bi￾modal, peaking near 0 and 1, corresponding to rock-dominated … view at source ↗
Figure 11
Figure 11. Figure 11: Distribution of observed planets and synthetic populations in the 3D logarithmic space defined by RHill, qp/s , and Rp. The colours and symbols follow the same definitions as in [PITH_FULL_IMAGE:figures/full_fig_p008_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Kernel density estimates illustrating the distributions of three formation-related parameters for planets associated with the four clus￾ters: HG (A1), WJD (A2), LMG (B), and VMGG (C). The parameters shown are the gas availability parameter G1, the gas mass fraction fgas, and the ice–rock mass ratio. For clarity, only synthetic planets within the overlap region of the observed parameter space (as defined i… view at source ↗
Figure 13
Figure 13. Figure 13: From top to bottom, the panels show synthetic populations mapped into the clustered parameter space in the logarithmic a–qp/s planes, with inner colours indicating different formation-related prop￾erties: G1 (Top), fgas (Middle), and ice–rock mass ratio (Bottom). In all panels, edge colours indicate the GMM cluster assignments (A1, A2, B, and C) inferred from data, while the colour bars show the correspon… view at source ↗
read the original abstract

Close-in exoplanets exhibit a wide range of orbital architectures and physical properties shaped by both formation conditions and migration processes. Although population-synthesis models predict distinct planetary populations, establishing a quantitative connection between observed exoplanets and synthetic populations remains challenging. We investigate the intrinsic organisation of close-in exoplanets using physically motivated dynamical parameters and connect the resulting populations to pebble-accretion formation pathways. A two-stage Gaussian mixture model (GMM) is applied to an observed sample of close-in exoplanets, performing unsupervised probabilistic clustering in a feature space dominated by dynamical descriptors of planet-star interactions. The resulting clusters are mapped onto a pebble-accretion synthetic population within a statistically motivated three-dimensional parameter space. Formation-related quantities, including gas availability, gas fraction, and ice-rock mass ratio, are then used to interpret the mapped populations. We identify statistically supported sub-populations without imposing predefined classification boundaries, including very-massive gas giants, hot giants, warm-Jupiter-dominated systems, and lower-mass giants. The mapped synthetic populations reveal systematic differences in formation timing, gas accretion, and solid growth histories. In particular, very-massive gas giants are preferentially associated with earlier formation epochs than hot-giant and warm-Jupiter-dominated populations. These results demonstrate that physically motivated machine-learning approaches can provide a statistically robust framework for linking observed exoplanet populations to theoretical planet formation pathways.

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 / 0 minor

Summary. The manuscript applies a two-stage Gaussian mixture model (GMM) to perform unsupervised clustering of close-in exoplanets in a feature space dominated by dynamical descriptors of planet-star interactions. The resulting clusters are mapped onto a pebble-accretion synthetic population in a three-dimensional parameter space (gas availability, gas fraction, ice-rock mass ratio) to interpret formation-related quantities and identify sub-populations (very-massive gas giants, hot giants, warm-Jupiter-dominated systems, lower-mass giants). The mapped populations are reported to show systematic differences in formation timing, gas accretion, and solid growth histories, with very-massive gas giants preferentially linked to earlier formation epochs than other populations.

Significance. If the cluster-to-synthetic mapping can be shown to isolate formation pathways independently of post-formation migration and detection biases, the work would supply a statistically motivated framework for connecting observed dynamical architectures to pebble-accretion theory. The approach is novel in its use of physically motivated dynamical features for unsupervised clustering, but its significance hinges on whether the reported timing differences survive explicit forward-modeling of Type-II migration and radial-velocity/transit selection effects.

major comments (2)
  1. The central mapping of GMM clusters onto the pebble-accretion synthetic population in the 3D space of gas availability, gas fraction, and ice-rock mass ratio is not demonstrated to be robust against Type-II migration and disk-driven evolution, which are known to modify the dynamical descriptors used for clustering. Without an explicit forward-modeling test of these post-formation effects on the observed sample, the reported systematic differences in formation timing (very-massive gas giants tied to earlier epochs) cannot be isolated from the alternative that clusters primarily reflect survival and observability filters.
  2. No sample size, cluster-validation metrics (e.g., BIC, silhouette scores, or stability under bootstrap), or error analysis on the two-stage GMM is supplied, even though the abstract asserts that the sub-populations are statistically supported. This information is load-bearing for the claim that the clusters reveal distinct formation pathways rather than being shaped by the feature-space construction or normalization choices.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed report. The comments highlight important limitations in our current analysis, and we address each point below with plans for revision.

read point-by-point responses

  1. Referee: [—] The central mapping of GMM clusters onto the pebble-accretion synthetic population in the 3D space of gas availability, gas fraction, and ice-rock mass ratio is not demonstrated to be robust against Type-II migration and disk-driven evolution, which are known to modify the dynamical descriptors used for clustering. Without an explicit forward-modeling test of these post-formation effects on the observed sample, the reported systematic differences in formation timing (very-massive gas giants tied to earlier epochs) cannot be isolated from the alternative that clusters primarily reflect survival and observability filters.

    Authors: We agree this is a substantive limitation. The clustering relies on observed dynamical features that can be altered by Type-II migration, and our mapping to the synthetic population does not include explicit forward modeling to isolate formation timing from post-formation effects. In revision we will add a dedicated limitations subsection discussing how migration and selection biases could influence the dynamical descriptors and the resulting formation-epoch interpretations. We will temper the claims to emphasize that the reported differences are suggestive links rather than fully isolated formation pathways, and note that a full forward-modeling test lies beyond the present scope. revision: yes

  2. Referee: [—] No sample size, cluster-validation metrics (e.g., BIC, silhouette scores, or stability under bootstrap), or error analysis on the two-stage GMM is supplied, even though the abstract asserts that the sub-populations are statistically supported. This information is load-bearing for the claim that the clusters reveal distinct formation pathways rather than being shaped by the feature-space construction or normalization choices.

    Authors: We accept that these quantitative details are required to support the statistical claims. The revised manuscript will report the exact sample size of the close-in exoplanet catalog, the BIC and silhouette scores used to select the number of GMM components in each stage, bootstrap resampling results for cluster stability, and uncertainty estimates on cluster assignments and mapped parameters. These will be presented in a new methods subsection on clustering validation. revision: yes

Circularity Check

0 steps flagged

No circularity: unsupervised clustering on observed dynamics mapped to independent synthetics

full rationale

The derivation applies a two-stage GMM to observed close-in exoplanets using dynamical descriptors of planet-star interactions, then maps the resulting clusters onto a separate pebble-accretion synthetic population in a 3D parameter space of gas availability, gas fraction and ice-rock mass ratio. No quoted step defines the clusters or the mapping in terms of the formation-timing outcomes being interpreted, nor does any prediction reduce by construction to a fitted parameter or self-citation chain. The chain from data-driven clustering to interpretation of synthetic populations remains independent of the target claims.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No free parameters, axioms, or invented entities can be identified from the abstract alone; the GMM number of components and the choice of dynamical feature space are standard but unspecified here.

pith-pipeline@v0.9.1-grok · 5791 in / 1207 out tokens · 23779 ms · 2026-06-27T08:35:20.773957+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

59 extracted references · 51 canonical work pages · 16 internal anchors

  1. [1]

    A new metric to quantify the similarity between planetary systems - application to dimensionality reduction using T-SNE

    New metric to quantify the similarity between planetary systems: application to dimensionality reduction using T-SNE. , keywords =. doi:10.1051/0004-6361/201834592 , archivePrefix =. 1901.09719 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201834592 1901

  2. [2]

    Using Deep Neural Networks to compute the mass of forming planets

    Using deep neural networks to compute the mass of forming planets. , keywords =. doi:10.1051/0004-6361/201834942 , archivePrefix =. 1903.00320 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201834942 1903

  3. [3]

    , keywords =

    A transformer-based generative model for planetary systems. , keywords =. doi:10.1051/0004-6361/202452297 , archivePrefix =. 2509.07226 , primaryClass =

    work page doi:10.1051/0004-6361/202452297

  4. [4]

    Machine learning classification of variable stars and eclipsing binaries in K2 fields 0–4 , journal =

    K2 variable catalogue – II. Machine learning classification of variable stars and eclipsing binaries in K2 fields 0–4 , journal =. 2015 , month =. doi:10.1093/mnras/stv2836 , url =

    work page doi:10.1093/mnras/stv2836 2015

  5. [5]

    , keywords =

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

    work page doi:10.1093/mnras/staa2498 2008

  6. [6]

    The growth of planets by pebble accretion in evolving protoplanetary discs

    The growth of planets by pebble accretion in evolving protoplanetary discs. , keywords =. doi:10.1051/0004-6361/201526463 , archivePrefix =. 1507.05209 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201526463

  7. [7]

    Pebble isolation mass --- scaling law and implications for the formation of super-Earths and gas giants

    Pebble-isolation mass: Scaling law and implications for the formation of super-Earths and gas giants , journal =. doi:10.1051/0004-6361/201731931 , archivePrefix =. 1801.02341 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201731931

  8. [8]

    2023 , volume =

    Hands-On Equations Balance Model Enhances Algebraic Equation Solving in Upper Elementary and Middle School Students , journal =. 2023 , volume =

    2023

  9. [9]

    NGTS-26 b and NGTS-27 b , journal =

    NGTS discovery of a highly inflated Saturn-mass planet and a highly irradiated hot Jupiter. NGTS-26 b and NGTS-27 b , journal =. doi:10.1051/0004-6361/202347162 , adsurl =

    work page doi:10.1051/0004-6361/202347162

  10. [10]

    arXiv e-prints , keywords =

    Exoplanet formation inference using conditional invertible neural networks. arXiv e-prints , keywords =. doi:10.48550/arXiv.2512.05751 , archivePrefix =. 2512.05751 , primaryClass =

    work page doi:10.48550/arxiv.2512.05751

  11. [11]

    Inside-Out Planet Formation

    Inside-out Planet Formation. , keywords =. doi:10.1088/0004-637X/780/1/53 , archivePrefix =. 1306.0576 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/0004-637x/780/1/53

  12. [12]

    Vulcan Planets: Inside-Out Formation of the Innermost Super-Earths

    Vulcan Planets: Inside-out Formation of the Innermost Super-Earths. , keywords =. doi:10.1088/2041-8205/798/2/L32 , archivePrefix =. 1411.2629 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1088/2041-8205/798/2/l32 2041

  13. [13]

    2023 , doi =

    The evolution of hot Jupiters revealed by the age distribution of their host stars , journal =. 2023 , doi =

    2023

  14. [14]

    I., & Johnson, J

    Origins of Hot Jupiters , journal =. doi:10.1146/annurev-astro-081817-051853 , archivePrefix =. 1801.06117 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1146/annurev-astro-081817-051853

  15. [15]

    The New Generation Planetary Population Synthesis (NGPPS). I. Bern global model of planet formation and evolution, model tests, and emerging planetary systems , journal =. doi:10.1051/0004-6361/202038553 , archivePrefix =. 2007.05561 , primaryClass =

    work page doi:10.1051/0004-6361/202038553 2007

  16. [16]

    , volume =

    Feinstein, C and Baume, G and Vergne, M M and Rodríguez, M J , title =. , volume =. 2024 , month =. doi:10.1093/mnras/stae2028 , url =

    work page doi:10.1093/mnras/stae2028 2024

  17. [17]

    Architectures of Planetary Systems and Implications for their Formation

    Architectures of planetary systems and implications for their formation. PNAS , keywords =. doi:10.1073/pnas.1304219111 , archivePrefix =. 1404.3157 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1073/pnas.1304219111

  18. [18]

    2024 , issn =

    A review of unsupervised learning in astronomy , journal =. 2024 , issn =. doi:https://doi.org/10.1016/j.ascom.2024.100851 , url =

    work page doi:10.1016/j.ascom.2024.100851 2024

  19. [19]

    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

  20. [20]

    , keywords =

    The dynamics of planetary rings. , keywords =. doi:10.1146/annurev.aa.20.090182.001341 , adsurl =

    work page doi:10.1146/annurev.aa.20.090182.001341

  21. [21]

    N., Helled, R., et al

    Giant Planets from the Inside-Out. Protostars and Planets VII , year = 2023, editor =. doi:10.48550/arXiv.2205.04100 , archivePrefix =. 2205.04100 , primaryClass =

    work page doi:10.48550/arxiv.2205.04100 2023

  22. [22]

    F., Millholland, S

    A hot-Jupiter progenitor on a super-eccentric retrograde orbit. , year = 2024, month = aug, volume =. doi:10.1038/s41586-024-07688-3 , adsurl =

    work page doi:10.1038/s41586-024-07688-3 2024

  23. [23]

    2020 , month =

    Optimizing exoplanet atmosphere retrieval using unsupervised machine-learning classification , journal =. 2020 , month =. doi:10.1093/mnras/staa978 , url =

    work page doi:10.1093/mnras/staa978 2020

  24. [24]

    doi:10.3847/1538-3881/ae1ba3 , adsurl =

    Detection of Exoplanets with Machine Learning Techniques Through Transit Light-curve Analysis , journal =. doi:10.3847/1538-3881/ae1ba3 , adsurl =

    work page doi:10.3847/1538-3881/ae1ba3

  25. [25]

    Inside-out Planet Formation. III. Planet-Disk Interaction at the Dead Zone Inner Boundary. , keywords =. doi:10.3847/0004-637X/816/1/19 , archivePrefix =. 1508.02791 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3847/0004-637x/816/1/19

  26. [26]

    Hot super-Earth systems from breaking compact resonant chains

    Formation of planetary systems by pebble accretion and migration. Hot super-Earth systems from breaking compact resonant chains. , keywords =. doi:10.1051/0004-6361/201935336 , archivePrefix =. 1902.08772 , primaryClass =

    work page doi:10.1051/0004-6361/201935336 1902

  27. [27]

    MNRAS , volume =

    Jiang, Ing-Guey and Yeh, Li-Chin and Hung, Wen-Liang and Yang, Miin-Shen , title =. MNRAS , volume =. 2006 , month =

    2006

  28. [28]

    Forming Planets via Pebble Accretion. Annu. Rev. Earth Planet. Sci. , year = 2017, month = aug, volume =. doi:10.1146/annurev-earth-063016-020226 , adsurl =

    work page doi:10.1146/annurev-earth-063016-020226 2017

  29. [29]

    How planetary growth outperforms migration

    How planetary growth outperforms migration , journal =. doi:10.1051/0004-6361/201834071 , archivePrefix =. 1811.00523 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201834071

  30. [30]

    Science Advances , keywords =

    A pebble accretion model for the formation of the terrestrial planets in the Solar System. Science Advances , keywords =. doi:10.1126/sciadv.abc0444 , archivePrefix =. 2102.08611 , primaryClass =

    work page doi:10.1126/sciadv.abc0444

  31. [31]

    , keywords =

    Exploring the conditions for forming planetesimals by the streaming instability and planetary systems by pebble accretion. , keywords =

  32. [32]

    2025 , publisher =

    Machine Learning for Exoplanet Detection: A Comparative Analysis using Kepler Data , journal =. 2025 , publisher =. doi:10.22128/ijaa.2025.2996.1219 , keywords =

    work page doi:10.22128/ijaa.2025.2996.1219 2025

  33. [33]

    doi:10.33232/001c.124538 , archivePrefix =

    Estimating Exoplanet Mass using Machine Learning on Incomplete Datasets , journal =. doi:10.33232/001c.124538 , archivePrefix =. 2410.06922 , primaryClass =

    work page doi:10.33232/001c.124538

  34. [34]

    Forming the cores of giant planets from the radial pebble flux in protoplanetary discs

    Forming the cores of giant planets from the radial pebble flux in protoplanetary discs. , keywords =. doi:10.1051/0004-6361/201424343 , archivePrefix =. 1408.6094 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201424343

  35. [35]

    , title =

    Mamajek, Eric E. , title =. AIP Conf. Proc. , volume =. 2009 , month =. doi:10.1063/1.3215910 , url =

    work page doi:10.1063/1.3215910 2009

  36. [36]

    , year = 1995, month = nov, volume =

    A Jupiter-mass companion to a solar-type star. , year = 1995, month = nov, volume =. doi:10.1038/378355a0 , adsurl =

    work page doi:10.1038/378355a0 1995

  37. [37]

    Characterization of exoplanets from their formation. I. Models of combined planet formation and evolution , journal =. doi:10.1051/0004-6361/201118457 , archivePrefix =. 1206.6103 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1051/0004-6361/201118457

  38. [38]

    2023 , month =

    Revisiting mass–radius relationships for exoplanet populations: a machine learning insight , journal =. 2023 , month =. doi:10.1093/mnras/stad2506 , url =

    work page doi:10.1093/mnras/stad2506 2023

  39. [39]

    doi:10.1051/0004-6361/202348690 , archivePrefix =

    The mass-radius relation of exoplanets revisited , journal =. doi:10.1051/0004-6361/202348690 , archivePrefix =. 2311.12593 , primaryClass =

    work page doi:10.1051/0004-6361/202348690

  40. [40]

    Planet population synthesis driven by pebble accretion in cluster environments

    Planet population synthesis driven by pebble accretion in cluster environments. , keywords =. doi:10.1093/mnras/stx2815 , archivePrefix =. 1710.10863 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1093/mnras/stx2815

  41. [41]

    Turkish Physical Society 33rd International Physics Congress (TPS33) , year = 2018, series =

    On the classification of exoplanets according to Safronov number. Turkish Physical Society 33rd International Physics Congress (TPS33) , year = 2018, series =. doi:10.1063/1.5025961 , adsurl =

    work page doi:10.1063/1.5025961 2018

  42. [42]

    2023 , month =

    Stable lifetime of compact, evenly spaced planetary systems with non-equal masses , journal =. 2023 , month =. doi:10.1093/mnras/stad393 , url =

    work page doi:10.1093/mnras/stad393 2023

  43. [43]

    The Protoplanetary Cloud and Its Evolution , journal =

  44. [44]

    , keywords =

    Machine learning-based identification of Gaia astrometric exoplanet orbits. , keywords =. doi:10.1093/mnras/staf018 , archivePrefix =. 2404.09350 , primaryClass =

    work page doi:10.1093/mnras/staf018

  45. [45]

    2021 , month =

    Postulating exoplanetary habitability via a novel anomaly detection method , journal =. 2021 , month =. doi:10.1093/mnras/stab3556 , url =

    work page doi:10.1093/mnras/stab3556 2021

  46. [46]

    , volume =

    Schanche, N and Cameron, A Collier and Hébrard, G and Nielsen, L and Triaud, A H M J and Almenara, J M and Alsubai, K A and Anderson, D R and Armstrong, D J and Barros, S C C and Bouchy, F and Boumis, P and Brown, D J A and Faedi, F and Hay, K and Hebb, L and Kiefer, F and Mancini, L and Maxted, P F L and Palle, E and Pollacco, D L and Queloz, D and Small...

    work page doi:10.1093/mnras/sty3146 2018

  47. [47]

    The New Generation Planetary Population Synthesis (NGPPS). V. Predetermination of planet types in global core accretion models , journal =. doi:10.1051/0004-6361/202140551 , archivePrefix =. 2104.11750 , primaryClass =

    work page doi:10.1051/0004-6361/202140551

  48. [48]

    2025 , issn =

    Galaxy morphological classification with manifold learning , journal =. 2025 , issn =. doi:https://doi.org/10.1016/j.ascom.2025.100963 , url =

    work page doi:10.1016/j.ascom.2025.100963 2025

  49. [49]

    2019 , url =

    Machine Learning Pipeline for Exoplanet Classification , journal =. 2019 , url =

    2019

  50. [50]

    doi:10.1073/pnas.2001258117 , archivePrefix =

    Predicting the long-term stability of compact multiplanet systems , journal =. doi:10.1073/pnas.2001258117 , archivePrefix =. 2007.06521 , primaryClass =

    work page doi:10.1073/pnas.2001258117 2007

  51. [51]

    An Overview of Inside-Out Planet Formation

    An Overview of Inside-Out Planet Formation. IAU Focus Meeting , keywords =. doi:10.1017/S1743921316002313 , archivePrefix =. 1510.06703 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1017/s1743921316002313

  52. [52]

    2014 , volume =

    The Need to Report Effect Size Estimates Revisited: An Overview of Some Recommended Measures of Effect Size , journal =. 2014 , volume =

    2014

  53. [53]

    Viualizing data using t-SNE , volume =

    van der Maaten, Laurens and Hinton, Geoffrey and Rachmad, Yoesoep , year =. Viualizing data using t-SNE , volume =

  54. [54]

    The Occurrence and Architecture of Exoplanetary Systems

    The Occurrence and Architecture of Exoplanetary Systems. , keywords =. doi:10.1146/annurev-astro-082214-122246 , archivePrefix =. 1410.4199 , primaryClass =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.1146/annurev-astro-082214-122246

  55. [55]

    doi:10.1038/s41586-020-2800-0 , archivePrefix =

    Stellar clustering shapes the architecture of planetary systems , journal =. doi:10.1038/s41586-020-2800-0 , archivePrefix =. 2010.10531 , primaryClass =

    work page doi:10.1038/s41586-020-2800-0 2010

  56. [56]

    2019 , month =

    Dynamical instability and its implications for planetary system architecture , journal =. 2019 , month =. doi:10.1093/mnras/stz054 , url =

    work page doi:10.1093/mnras/stz054 2019

  57. [57]

    2008 , issn =

    Learning a Mahalanobis distance metric for data clustering and classification , journal =. 2008 , issn =. doi:https://doi.org/10.1016/j.patcog.2008.05.018 , author =

    work page doi:10.1016/j.patcog.2008.05.018 2008

  58. [58]

    doi:10.3847/1538-4357/adddb3 , archivePrefix =

    The Narrow Formation Pathway of Hot Saturns: Constraints on Initial Planetary Properties , journal =. doi:10.3847/1538-4357/adddb3 , archivePrefix =. 2505.23148 , primaryClass =

    work page doi:10.3847/1538-4357/adddb3

  59. [59]

    , keywords =

    Improving Earth-like planet detection in radial velocity using deep learning. , keywords =. doi:10.1051/0004-6361/202450022 , archivePrefix =. 2405.13247 , primaryClass =

    work page doi:10.1051/0004-6361/202450022