Pith. sign in

REVIEW 2 major objections 2 minor 66 references

A transformer surrogate trained on multi-fluid MHD outputs predicts Europa magnetic perturbations at 40,000 times the speed of the parent code while matching its accuracy.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.3

2026-06-27 13:59 UTC pith:SXZS63VC

load-bearing objection LEAP gives a practical 40,000x speed-up for Europa MHD surrogates but the abstract leaves generalization to new plasma conditions untested. the 2 major comments →

arxiv 2606.10215 v1 pith:SXZS63VC submitted 2026-06-08 physics.space-ph physics.plasm-ph

LEAP: A Rapid Neural Surrogate of Multi-Fluid MHD at Europa

Sachin Alexander Reddy , Abigail R. Azari , Corey J. Cochrane , Xianzhe Jia , Tom A. Nordheim , Lukas Mandrake , Steven D. Vance , Camilla Harris
show 1 more author
Ioana Ciuc\v{a}
This is my paper
classification physics.space-ph physics.plasm-ph
keywords Europamulti-fluid MHDneural surrogatemagnetic field perturbationsplasma interactiontransformer modelEuropa ClipperJUICE
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper shows that a neural network can be trained on outputs from a full multi-fluid MHD simulation of Jupiter's plasma interacting with Europa to predict the resulting magnetic field perturbations along any spacecraft trajectory. Because the surrogate runs in milliseconds on ordinary hardware, it removes the computational barrier that has limited how many different plasma conditions and geometries can be explored. A sympathetic reader would see this as opening the door to systematic surveys and statistical inference on the magnetic signatures that the upcoming Europa Clipper and JUICE missions will measure, signatures that carry information about the moon's subsurface ocean. The reported test-set error of roughly 2.6 nT and the close match to the parent MHD model on the Galileo E4 and E14 flybys are presented as evidence that the learned mapping is faithful enough for these new uses.

Core claim

LEAP is a transformer-based model trained directly on the output fields of a state-of-the-art multi-fluid MHD code; once trained it reproduces the magnetic perturbations that the MHD code would produce along any given trajectory, achieving test errors of ±2.6 nT and the same accuracy as the parent model on the E4 and E14 Galileo encounters, all while evaluating in milliseconds rather than the twelve hours required by the original simulation.

What carries the argument

Transformer-based neural surrogate trained on multi-fluid MHD simulation outputs to map plasma parameters and spacecraft trajectories to magnetic field perturbations.

Load-bearing premise

The set of MHD runs used for training already covers the full range of plasma conditions and flyby geometries that real Europa encounters will produce.

What would settle it

A new multi-fluid MHD run or an actual spacecraft magnetic-field time series, drawn from conditions outside the training distribution, in which the LEAP predictions deviate from the MHD results by substantially more than 2.6 nT.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Large-scale parameter surveys of plasma conditions become feasible on ordinary computers.
  • Probabilistic estimates of unknown plasma parameters can be obtained by repeated fast forward evaluations.
  • The surrogate can be used to propose new MHD runs that fill gaps in the training distribution.
  • The same training approach can be applied to other bodies once suitable MHD data exist.

Where Pith is reading between the lines

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

  • Mission planning teams could run thousands of forward models during real-time operations to test different trajectory or instrument configurations.
  • The framework could be retrained periodically as new MHD data become available, gradually improving coverage without discarding prior work.
  • Similar surrogates might be trained for the induced fields of other icy moons where plasma interactions complicate ocean inference.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 2 minor

Summary. The paper introduces LEAP, a transformer-based neural surrogate trained on outputs from a multi-fluid MHD code to predict magnetic field perturbations along spacecraft trajectories at Europa. It reports test-set errors of ±2.6 nT, matches the parent MHD model on the Galileo E4 and E14 flybys, and achieves a ~40,000x speedup (milliseconds on a laptop vs. 12 hours on HPC), enabling large-scale parameter surveys and probabilistic inference of plasma conditions.

Significance. If the reported accuracy generalizes reliably, the work provides a practical tool for accelerating plasma-interaction modeling at icy moons, directly supporting data interpretation for Europa Clipper and JUICE by making extensive parameter explorations computationally feasible. The explicit speed-up quantification and direct comparison to MHD on real flybys are concrete strengths.

major comments (2)
  1. [Abstract] Abstract: the central claim that LEAP 'matches the parent MHD model in accuracy' on E4/E14 and supports 'large-scale parameter surveys' requires evidence that the two flybys were held out of training and that the training ensemble spans the relevant ranges of upstream density, velocity, temperature, and magnetic-field orientation. No such information is provided, leaving the out-of-distribution generalization unverified.
  2. [Results] Results section (implied by abstract claims): the reported test-set error of ±2.6 nT is given without details on the training/validation split, hyperparameter selection, or any explicit out-of-distribution tests on trajectories or plasma parameters outside the training distribution. This directly undermines the generalization required for the probabilistic-estimation use case.
minor comments (2)
  1. [Abstract] Abstract: the phrase 'matches the parent MHD model in accuracy' would be clearer if it quantified the point-wise or integrated difference (e.g., RMS or maximum deviation) rather than a binary statement.
  2. [Methods] The manuscript would benefit from a table or figure summarizing the span of plasma parameters used in the MHD training runs.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their constructive comments, which highlight the need for greater transparency on data provenance and generalization. We will revise the manuscript to address these points directly.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that LEAP 'matches the parent MHD model in accuracy' on E4/E14 and supports 'large-scale parameter surveys' requires evidence that the two flybys were held out of training and that the training ensemble spans the relevant ranges of upstream density, velocity, temperature, and magnetic-field orientation. No such information is provided, leaving the out-of-distribution generalization unverified.

    Authors: We agree that the abstract claims require supporting details that are not currently stated. In revision we will expand the abstract and add a new subsection (Methods or Results) that explicitly reports: (i) the ranges of upstream density, velocity, temperature, and magnetic-field orientation used to generate the training ensemble, (ii) the train/validation/test split, and (iii) confirmation that the Galileo E4 and E14 trajectories were excluded from training. These additions will directly substantiate the generalization claims and the suitability for parameter surveys. revision: yes

  2. Referee: [Results] Results section (implied by abstract claims): the reported test-set error of ±2.6 nT is given without details on the training/validation split, hyperparameter selection, or any explicit out-of-distribution tests on trajectories or plasma parameters outside the training distribution. This directly undermines the generalization required for the probabilistic-estimation use case.

    Authors: We concur that the current Results section lacks these methodological details. We will revise it to include: (1) the precise training/validation/test split ratios and how the test set was constructed, (2) the hyperparameter selection procedure (including any cross-validation or search strategy employed), and (3) explicit out-of-distribution evaluations on both unseen trajectories and plasma-parameter combinations lying outside the training distribution. These additions will strengthen support for the reported test error and the probabilistic-inference use case. revision: yes

Circularity Check

0 steps flagged

No circularity: surrogate trained on independent MHD outputs with held-out test evaluation

full rationale

The paper presents LEAP as a transformer surrogate trained on outputs from an external multi-fluid MHD code. Reported metrics are test-set errors of ±2.6 nT and agreement on Galileo E4/E14 flybys, which are standard held-out evaluations rather than quantities defined by construction from fitted parameters inside the paper. No self-definitional loops, fitted-input predictions, or load-bearing self-citations appear in the abstract or described claims; the central result (speed-up with comparable accuracy) rests on independent simulation data and does not reduce to renaming or ansatz smuggling. Generalization concerns are separate from circularity.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

Central claim rests on the domain assumption that the parent multi-fluid MHD code is an adequate ground truth and that its outputs span the relevant input space for Europa encounters. No free parameters are explicitly fitted beyond standard neural network training; no new physical entities are introduced.

free parameters (1)
  • transformer model weights
    Learned during supervised training on MHD simulation outputs; these are the fitted parameters that determine the surrogate mapping.
axioms (1)
  • domain assumption Multi-fluid MHD simulations provide a sufficiently accurate representation of Jupiter-Europa plasma interactions for the purpose of surrogate training.
    The surrogate is defined to reproduce MHD outputs; any systematic bias in the MHD model propagates directly to LEAP predictions.

pith-pipeline@v0.9.1-grok · 5796 in / 1401 out tokens · 21798 ms · 2026-06-27T13:59:31.382968+00:00 · methodology

0 comments
read the original abstract

Characterizing Europa's subsurface ocean is a key objective of the Europa Clipper and JUICE missions in the search for life beyond Earth. Although the ocean's induced magnetic field provides key constraints on habitability, interpretation is complicated by perturbations arising from Jupiter's plasma interaction with Europa. Physics-based models (e.g. magnetohydrodynamic, MHD) required to characterize these effects are physically comprehensive, but have a prohibitive computational cost. To address this, we introduce Learning Europa's Atmosphere and Plasma (LEAP), a transformer-based surrogate trained on outputs from a state-of-the-art multi-fluid MHD code to predict magnetic field perturbations along spacecraft trajectories. LEAP evaluates in milliseconds on a laptop, whereas MHD takes 12 hrs on a high-performance computer (~40,000x speed-up). The model has test set errors of -/+ 2.6 nT, and for the Galileo E4 and E14 flybys of Europa it matches the parent MHD model in accuracy. Its enhanced speed enables large-scale parameter surveys and probabilistic estimations of plasma conditions, establishing a new framework for accelerated plasma interaction modeling. LEAP can also inform future MHD simulations while learning from them. Beyond Europa, this framework could be expanded to planning future missions or to other high-priority bodies, including Uranus and Neptune.

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

66 extracted references · 4 canonical work pages · 1 internal anchor

  1. [1]

    D.et al.Investigating europa’s habitability with the europa clipper

    Vance, S. D.et al.Investigating europa’s habitability with the europa clipper. Space Science Reviews219, 81 (2023)

    2023

  2. [2]

    Grasset, O.et al.Jupiter icy moons explorer (juice): An esa mission to orbit ganymede and to characterise the jupiter system.Planetary and Space Science 78, 1–21 (2013)

    2013

  3. [3]

    T.et al.Science overview of the europa clipper mission.Space Science Reviews220, 40 (2024)

    Pappalardo, R. T.et al.Science overview of the europa clipper mission.Space Science Reviews220, 40 (2024)

    2024

  4. [4]

    Khurana, K.et al.Induced magnetic fields as evidence for subsurface oceans in europa and callisto.Nature395, 777–780 (1998)

    1998

  5. [5]

    G.et al.Galileo magnetometer measurements: A stronger case for a subsurface ocean at europa.Science289, 1340–1343 (2000)

    Kivelson, M. G.et al.Galileo magnetometer measurements: A stronger case for a subsurface ocean at europa.Science289, 1340–1343 (2000)

    2000

  6. [6]

    Zimmer, C., Khurana, K. K. & Kivelson, M. G. Subsurface oceans on europa and callisto: Constraints from galileo magnetometer observations.Icarus147, 329–347 (2000)

    2000

  7. [7]

    G., Khurana, K

    Jia, X., Kivelson, M. G., Khurana, K. K. & Walker, R. J. Magnetic fields of the satellites of jupiter and saturn.Space Science Reviews152, 271–305 (2010)

    2010

  8. [8]

    G.et al.Magnetospheric interactions with satellites.Jupiter: The planet, satellites and magnetosphere1, 513–536 (2004)

    Kivelson, M. G.et al.Magnetospheric interactions with satellites.Jupiter: The planet, satellites and magnetosphere1, 513–536 (2004)

    2004

  9. [9]

    Neubauer, F. M. Alfv´ en wings and electromagnetic induction in the interiors: Europa and callisto.Journal of Geophysical Research: Space Physics104, 28671– 28684 (1999)

    1999

  10. [10]

    & Kivelson, M

    Volwerk, M., Khurana, K. & Kivelson, M. Europa’s alfv´ en wing: Shrinkage and displacement influenced by an induced magnetic field.Annales Geophysicae25, 905–914 (2007). 26

    2007

  11. [11]

    Harris, C. D.et al.Multi-fluid mhd simulations of europa’s plasma interac- tion under different magnetospheric conditions.Journal of Geophysical Research: Space Physics126, e2020JA028888 (2021)

    2021

  12. [12]

    J., Vance, S

    Cochrane, C. J., Vance, S. D., Castillo-Rogez, J. C., Styczinski, M. J. & Liuzzo, L. Stronger evidence of a subsurface ocean within callisto from a multifrequency investigation of its induced magnetic field.AGU Advances6, e2024AV001237 (2025)

    2025

  13. [13]

    Nonlinear standing alfv´ en wave current system at io: Theory

    Neubauer, F. Nonlinear standing alfv´ en wave current system at io: Theory. Journal of Geophysical Research: Space Physics85, 1171–1178 (1980)

    1980

  14. [14]

    Saur, J., Neubauer, F. M. & Glassmeier, K.-H. Induced magnetic fields in solar system bodies.Space science reviews152, 391–421 (2010)

    2010

  15. [15]

    G.et al.The europa clipper magnetometer.Space Science Reviews 219, 48 (2023)

    Kivelson, M. G.et al.The europa clipper magnetometer.Space Science Reviews 219, 48 (2023)

    2023

  16. [16]

    Kabin, K.et al.On europa’s magnetospheric interaction: A mhd simulation of the e4 flyby.Journal of Geophysical Research: Space Physics104, 19983–19992 (1999)

    1999

  17. [17]

    & Saur, J

    Schilling, N., Neubauer, F. & Saur, J. Influence of the internally induced magnetic field on the plasma interaction of europa.Journal of Geophysical Research: Space Physics113(2008)

    2008

  18. [18]

    Rubin, M.et al.Self-consistent multifluid mhd simulations of europa’s exospheric interaction with jupiter’s magnetosphere.Journal of Geophysical Research: Space Physics120, 3503–3524 (2015)

    2015

  19. [19]

    G., Khurana, K

    Jia, X., Kivelson, M. G., Khurana, K. K. & Kurth, W. S. Evidence of a plume on europa from galileo magnetic and plasma wave signatures.Nature Astronomy 2, 459–464 (2018)

    2018

  20. [20]

    D., Jia, X

    Harris, C. D., Jia, X. & Slavin, J. A. Multi-fluid mhd simulations of europa’s plasma interaction: Effects of variation in europa’s atmosphere.Journal of Geophysical Research: Space Physics127, e2022JA030569 (2022)

    2022

  21. [21]

    Allegrini, F.et al.Electron beams at europa.Geophysical research letters51, e2024GL108422 (2024)

    2024

  22. [22]

    Cervantes, S.et al.Mhd simulations of europa’s interaction with jupiter’s mag- netosphere during the juno flyby: Electron beams in the plasma wake.Journal of Geophysical Research: Space Physics130, e2025JA033825 (2025)

    2025

  23. [23]

    Computer Physics Communications182, 946–966 (2011)

    M¨ uller, J.et al.Aikef: Adaptive hybrid model for space plasma simulations. Computer Physics Communications182, 946–966 (2011). 27

    2011

  24. [24]

    M., Stahl, A

    Addison, P., Haynes, C. M., Stahl, A. M., Liuzzo, L. & Simon, S. Magnetic signatures of the interaction between europa and jupiter’s magnetosphere during the juno flyby.Geophysical Research Letters51, e2023GL106810 (2024)

    2024

  25. [25]

    Proceedings of the National Academy of Sciences , year =

    Cranmer, K., Brehmer, J. & Louppe, G. The frontier of simulation-based infer- ence.Proceedings of the National Academy of Sciences117, 30055–30062 (2020). URL https://www.pnas.org/doi/abs/10.1073/pnas.1912789117

    work page doi:10.1073/pnas.1912789117 2020

  26. [26]

    Vaswani, A.et al.Attention is all you need.Advances in neural information processing systems30(2017)

    2017

  27. [27]

    & Thuerey, N

    Holzschuh, B., Liu, Q., Kohl, G. & Thuerey, N. Singh, A.et al.(eds)PDE- transformer: Efficient and versatile transformers for physics simulations. (eds Singh, A.et al.)Proceedings of the 42nd International Conference on Machine Learning, Vol. 267 ofProceedings of Machine Learning Research, 23562–23602 (PMLR, 2025). URL https://proceedings.mlr.press/v267/h...

    2025

  28. [28]

    & Sun, H

    Wan, H., Zhang, R., Wang, Q., Liu, Y. & Sun, H. Pesanet: Physics-encoded spectral attention network for simulating pde-governed complex systems (2025). URL https://arxiv.org/abs/2505.01736. arXiv:2505.01736

    work page arXiv 2025

  29. [29]

    Duling, S.et al.Ganymede mhd model: Magnetospheric context for juno’s pj34 flyby.Geophysical Research Letters49, e2022GL101688 (2022)

    2022

  30. [30]

    R.et al.Simulation based inference for mission planning and parameter estimation (2023)

    Azari, A. R.et al.Simulation based inference for mission planning and parameter estimation (2023)

    2023

  31. [31]

    Tanaka, T. Finite volume tvd scheme on an unstructured grid system for three-dimensional mhd simulation of inhomogeneous systems including strong background potential fields.Journal of Computational Physics111, 381–389 (1994)

    1994

  32. [32]

    & Fujita, S

    Kataoka, R., Nakamizo, A., Nakano, S. & Fujita, S. Machine learning-based emu- lator for the physics-based simulation of auroral current system.Space Weather 22, e2023SW003720 (2024)

    2024

  33. [33]

    Holmberg, D.et al.Graph-based neural space weather forecasting (2025)

    2025

  34. [34]

    Von Alfthan, S.et al.Vlasiator: First global hybrid-vlasov simulations of earth’s foreshock and magnetosheath.Journal of Atmospheric and Solar-Terrestrial Physics120, 24–35 (2014)

    2014

  35. [35]

    The Astrophysical Journal905, 14 (2020)

    Burkhart, B.et al.The catalogue for astrophysical turbulence simulations (cats). The Astrophysical Journal905, 14 (2020)

    2020

  36. [36]

    Khurana, K. K. Euler potential models of jupiter’s magnetospheric field.Journal of Geophysical Research: Space Physics102, 11295–11306 (1997). 28

    1997

  37. [37]

    J., Hinson, D

    Kliore, A. J., Hinson, D. P., Flasar, F. M., Nagy, A. F. & Cravens, T. E. The ionosphere of europa from galileo radio occultations.Science277, 355–358 (1997)

    1997

  38. [38]

    & Erd, C

    Boutonnet, A., Langevin, Y. & Erd, C. Designing the juice trajectory.Space Science Reviews220, 67 (2024)

    2024

  39. [39]

    Masters, A.et al.Magnetosphere and plasma science with the jupiter icy moons explorer.Space Science Reviews221, 24 (2025)

    2025

  40. [40]

    & Melwani Daswani, M

    Styczinski, M., Vance, S. & Melwani Daswani, M. Planetprofile: self-consistent interior structure modeling for ocean worlds and rocky dwarf planets in python. Earth and Space Science10, e2022EA002748 (2023)

    2023

  41. [41]

    TabTransformer: Tabular Data Modeling Using Contextual Embeddings

    Huang, X., Khetan, A., Cvitkovic, M. & Karnin, Z. Tabtransformer: Tabular data modeling using contextual embeddings.arXiv preprint arXiv:2012.06678(2020)

    work page internal anchor Pith review Pith/arXiv arXiv 2012

  42. [42]

    H.et al.The plasma instrument for magnetic sounding (pims) on the europa clipper mission.Space Science Reviews219, 62 (2023)

    Westlake, J. H.et al.The plasma instrument for magnetic sounding (pims) on the europa clipper mission.Space Science Reviews219, 62 (2023)

    2023

  43. [43]

    B.et al.Revealing the interior structure of icy moons with a bayesian approach to magnetic induction measurements.The Planetary Science Journal4, 62 (2023)

    Biersteker, J. B.et al.Revealing the interior structure of icy moons with a bayesian approach to magnetic induction measurements.The Planetary Science Journal4, 62 (2023)

    2023

  44. [44]

    Styczinski, M. J. & Cochrane, C. J. Planetmag: Software for evaluation of outer planet magnetic fields and corresponding excitations at their moons.Earth and Space Science11, e2024EA003552 (2024)

    2024

  45. [45]

    & Cervantes, S

    Winkenstern, J., Saur, J. & Cervantes, S. An inversion of magnetic field mea- surements to constrain the depth, thickness, and conductivity of europa’s ocean. Journal of Geophysical Research: Planets130, e2025JE009122 (2025)

    2025

  46. [46]

    V.et al.Dusk over dawn o2 asymmetry in europa’s near-surface atmosphere.Planetary and Space Science167, 23–32 (2019)

    Oza, A. V.et al.Dusk over dawn o2 asymmetry in europa’s near-surface atmosphere.Planetary and Space Science167, 23–32 (2019)

    2019

  47. [47]

    & Neubauer, F

    Saur, J., Strobel, D. & Neubauer, F. Interaction of the jovian magnetosphere with europa: Constraints on the neutral atmosphere.Journal of Geophysical Research: Planets103, 19947–19962 (1998)

    1998

  48. [48]

    Meitzler, R.et al.Investigating europa’s radiation environment with the europa clipper radiation monitor.Space Science Reviews219, 61 (2023)

    2023

  49. [49]

    Nordheim, T.et al.Magnetospheric ion bombardment of europa’s surface.The Planetary Science Journal3, 5 (2022)

    2022

  50. [50]

    Frazier, W.et al.Trident: the path to triton on a discovery budget (2020)

    2020

  51. [51]

    Girija, A. P. A flagship-class uranus orbiter and probe mission concept using aerocapture.Acta Astronautica202, 104–118 (2023). 29

    2023

  52. [52]

    Glocer, A.et al.Multifluid block-adaptive-tree solar wind roe-type upwind scheme: Magnetospheric composition and dynamics during geomagnetic storms—initial results.Journal of Geophysical Research: Space Physics114 (2009)

    2009

  53. [53]

    T´ oth, G.et al.Adaptive numerical algorithms in space weather modeling.Journal of Computational Physics231, 870–903 (2012)

    2012

  54. [54]

    I.et al.What sustained multi-disciplinary research can achieve: The space weather modeling framework.Journal of Space Weather and Space Climate 11, 42 (2021)

    Gombosi, T. I.et al.What sustained multi-disciplinary research can achieve: The space weather modeling framework.Journal of Space Weather and Space Climate 11, 42 (2021)

    2021

  55. [55]

    Plainaki, C.et al.Exospheric o2 densities at europa during different orbital phases.Planetary and Space Science88, 42–52 (2013)

    2013

  56. [56]

    K., Kivelson, M

    Khurana, K. K., Kivelson, M. G., Hand, K. P. & Russell, C. T. Electromagnetic induction from europa’s ocean and the deep interior.Europa572–586 (2009)

    2009

  57. [57]

    & Weiss, B

    Cochrane, C., Vance, S., Biersteker, J., Styczinski, M. & Weiss, B. On detecting and characterizing planetary oceans in the solar system using a distance-based ensemble modelling approach: application to the uranus system.Philosophical Transactions A382, 20240086 (2024)

    2024

  58. [58]

    J.et al.Magnetic field modeling and visualization of the europa clipper spacecraft.Space Science Reviews219, 34 (2023)

    Cochrane, C. J.et al.Magnetic field modeling and visualization of the europa clipper spacecraft.Space Science Reviews219, 34 (2023)

    2023

  59. [59]

    A.et al.Europa clipper mission design, mission plan, and navigation.Space Science Reviews221, 22 (2025)

    Cangahuala, L. A.et al.Europa clipper mission design, mission plan, and navigation.Space Science Reviews221, 22 (2025)

    2025

  60. [60]

    Adam, K. D. B. J.et al.A method for stochastic optimization.arXiv preprint arXiv:1412.69801412(2014)

    work page arXiv 2014

  61. [61]

    & Price, R

    Bayes, T. & Price, R. An essay towards solving a problem in the doctrine of chances. By the late Rev. Mr. Bayes, F. R. S. communicated by Mr. Price, in a letter to John Canton, A. M. F. R. S.Philosophical Transactions of the Royal Society of London53, 370–418 (1763)

  62. [62]

    K.Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan(Elsevier, 2015)

    Kruschke, J. K.Doing Bayesian Data Analysis: A Tutorial with R, JAGS, and Stan(Elsevier, 2015)

    2015

  63. [63]

    & Fonnesbeck, C

    Salvatier, J., Wiecki, T. & Fonnesbeck, C. Probabilistic programming in Python using PyMC3.PeerJ Computer Science2:e55 (2016)

    2016

  64. [64]

    Abril-Pla, O.et al.PyMC: a modern, and comprehensive probabilistic program- ming framework in Python.PeerJ Computer Science9, e1516 (2023)

    2023

  65. [65]

    Hoffman, M. D. & Gelman, A. The No-U-Turn Sampler: Adaptively Setting Path Lengths in Hamiltonian Monte Carlo.Journal of Machine Learning Research15, 30 1593–1623 (2011)

    2011

  66. [66]

    Gurnett, D.et al.Galileo plasma wave observations near europa.Geophysical research letters25, 237–240 (1998). 31

    1998