- Physical Review Journals
All JournalsPhysics Magazine

Physical Review Research

Reuse & Permissions

It is not necessary to obtain permission to reuse this article or its components as it is available under the terms of the Creative Commons Attribution 4.0 International license. This license permits unrestricted use, distribution, and reproduction in any medium, provided attribution to the author(s) and the published article's title, journal citation, and DOI are maintained. Please note that some figures may have been included with permission from other third parties. It is your responsibility to obtain the proper permission from the rights holder directly for these figures.

Export citation

Export citation

Choose format for download:

Download Citation
  • Open Access

Estimating orbital parameters of direct imaging exoplanet using neural networks

Bo Liang1,2,*, Chang Liu3,*, Hanlin Song1,4,*,†, Tianyu Zhao1,‡, Yuxiang Xu1,5, Zihao Xiao1,3, Manjia Liang1, Minghui Du1, Wei-Liang Qian6 et al.

Li-e Qiang1, Mingming Sun7, Peng Xu1,8,9,§, and Ziren Luo1,8

  • *These authors contributed equally to this work.
  • †Contact author: hanlin@stu.pku.edu.cn
  • ‡Contact author: zhaotianyu@imech.ac.cn
  • §Contact author: xupeng@imech.ac.cn
PDF

Phys. Rev. Research 8, 023341 – Published 25 June, 2026

DOI: https://doi.org/10.1103/ykfh-cdzk

Abstract

We propose a flow-matching Markov chain Monte Carlo (FM-MCMC) algorithm for estimating the orbital parameters of exoplanetary systems, especially for those in which only one exoplanet is involved. Compared to traditional methods that rely on random sampling within the Bayesian framework, our approach first leverages flow-matching posterior estimation to efficiently constrain the prior range of physical parameters and then employs MCMC to accurately infer the posterior distribution. For example, in the orbital parameter inference of β Pictoris b, our model achieved a substantial speedup while maintaining comparable accuracy—running 77.8 times faster than parallel tempered MCMC and 365.4 times faster than nested sampling. Moreover, our FM-MCMC method also attained the highest average log-likelihood among all approaches, demonstrating its superior sampling efficiency and accuracy. This highlights the scalability and efficiency of our approach, making it well suited for processing the massive datasets expected from future exoplanet surveys. Beyond astrophysics, our methodology establishes a versatile paradigm for synergizing deep generative models with traditional sampling, which can be adopted to tackle complex inference problems in other fields, such as cosmology, biomedical imaging, and particle physics.

View figure in article

Physics Subject Headings (PhySH)

Article Text

Supplemental Material

References (42)

  1. A. Wolszczan and D. A. Frail, A planetary system around the millisecond pulsar PSR1257+12, Nature (London) 355, 145 (1992).
  2. M. Mayor and D. Queloz, A Jupiter-mass companion to a solar-type star, Nature (London) 378, 355 (1995).
  3. NASA Exoplanet Archive, NASA Exoplanet Archive by the data 2025.07.09, 2025, https://science.nasa.gov/exoplanets/, accessed 5 September 2025.
  4. A. R. Howe, J. C. Becker, C. C. Stark, and F. C. Adams, Architecture classification for extrasolar planetary systems, Astron. J. 169, 149 (2025).
  5. T. D. Brandt, T. J. Dupuy, Y. Li, G. M. Brandt, Y. Zeng, D. Michalik, D. C. Bardalez Gagliuffi, and V. Raposo-Pulido, orvara: An efficient code to fit orbits using radial velocity, absolute, and/or relative astrometry, Astron. J. 162, 186 (2021).
  6. S. Blunt, J. J. Wang, I. Angelo, H. Ngo, D. Cody, R. J. De Rosa, J. R. Graham, L. Hirsch, V. Nagpal, E. L. Nielsen, L. Pearce, M. Rice, and R. Tejada, orbitize!: A comprehensive orbit-fitting software package for the high-contrast imaging community, Astron. J. 159, 89 (2020).
  7. J. N. Winn, The transiting exoplanet survey satellite, arXiv:2410.12905.
  8. A. Fortier, A. E. Simon, C. Broeg, G. Olofsson, A. Deline, T. G. Wilson, P. F. Maxted, A. Brandeker, A. C. Cameron, M. Beck, et al., Cheops in-flight performance—A comprehensive look at the first 3.5 yr of operations, Astron. Astrophys. 687, A302 (2024).
  9. E. Pancino, Gaia: The galaxy in six (and more) dimensions, Adv. Space Res. 65, 1 (2020).
  10. J.-L. Beuzit, A. Vigan, D. Mouillet, K. Dohlen, R. Gratton, A. Boccaletti, J.-F. Sauvage, H. M. Schmid, M. Langlois, C. Petit, et al., Sphere: The exoplanet imager for the very large telescope, Astron. Astrophys. 631, A155 (2019).
  11. H. Rauer, C. Aerts, J. Cabrera, M. Deleuil, A. Erikson, L. Gizon, M. Goupil, A. Heras, J. Lorenzo-Alvarez, F. Marliani, C. Martin-Garcia, J. M. Mas-Hesse, L. O’Rourke, H. Osborn, I. Pagano, and G. Piotto, The Plato mission, Experimental Astron. 59, 26 (2025).
  12. Q. Changeat and K. H. Yip, ESA-Ariel data challenge NeurIPS 2022: Introduction to exo-atmospheric studies and presentation of the Atmospheric Big Challenge (ABC) database, RAS Techniques and Instruments 2, 45 (2023).
  13. T. Zingales, G. Tinetti, I. Pillitteri, J. Leconte, G. Micela, and S. Sarkar, The Ariel mission reference sample, Exp. Astron. 46, 67 (2018).
  14. A. Boccaletti, G. Chauvin, D. Mouillet, O. Absil, F. Allard, S. Antoniucci, J.-C. Augereau, P. Barge, A. Baruffolo, J.-L. Baudino, et al., Sphere+: Imaging young Jupiters down to the snowline, arXiv:2003.05714 .
  15. V. P. Bailey, E. Bendek, B. Monacelli, C. Baker, G. Bedrosian, E. Cady, E. S. Douglas, T. Groff, S. R. Hildebrandt, N. J. Kasdin, et al., Nancy Grace Roman Space Telescope coronagraph instrument overview and status, Techniques and Instrumentation for Detection of Exoplanets XI 12680, 283 (2023).
  16. A.-L. Maire, L. Delrez, F. J. Pozuelos, J. Becker, N. Espinoza, J. Lillo-Box, A. Revol, O. Absil, E. Agol, J. M. Almenara, et al., Workshop summary: Exoplanet orbits and dynamics, Publ. Astron. Soc. Pac. 135, 106001 (2023).
  17. D. A. Fischer, A. W. Howard, G. P. Laughlin, B. Macintosh, S. Mahadevan, J. Sahlmann, and J. C. Yee, Exoplanet detection techniques, in Protostars and Planets VI, edited by H. Beuther, R. S. Klessen, C. P. Dullemond, and T. Henning (University of Arizona Press, Tucson, AZ, 2014), pp. 715–737.
  18. H. J. Deeg and J. A. Belmonte, Handbook of Exoplanets (2018).
  19. R. Abuter, M. Accardo, A. Amorim, N. Anugu, G. Avila, N. Azouaoui, M. Benisty, J.-P. Berger, N. Blind, H. Bonnet, et al., First light for gravity: Phase referencing optical interferometry for the very large telescope interferometer, Astron. Astrophys. 602, A94 (2017).
  20. P. A. Lightsey, C. Atkinson, M. Clampin, and L. D. Feinberg, James Webb Space Telescope: Large deployable cryogenic telescope in space, Opt. Eng. 51, 011003 (2012).
  21. T. Forveille, J. L. Beuzit, X. Delfosse, D. Segransan, F. Beck, M. Mayor, C. Perrier, A. Tokovini, and S. Udry, Accurate masses of very low mass stars: I Gl 570bc (0.6+0.4 Msol), Astron. astrophys. 351, 619 (1999).
  22. S. Blunt, E. L. Nielsen, R. J. De Rosa, Q. M. Konopacky, D. Ryan, J. J. Wang, L. Pueyo, J. Rameau, C. Marois, F. Marchis, B. Macintosh, J. R. Graham, G. Duchêne, and A. C. Schneider, Orbits for the impatient: A Bayesian rejection-sampling method for quickly fitting the orbits of long-period exoplanets, Astron. J. 153, 229 (2017).
  23. D. Foreman-Mackey, D. W. Hogg, D. Lang, and J. Goodman, emcee: The MCMC hammer, Publ. Astron. Soc. Pac. 125, 306 (2013).
  24. W. D. Vousden, W. M. Farr, and I. Mandel, Dynamic temperature selection for parallel tempering in Markov chain Monte Carlo simulations, Mon. Not. R. Astron. Soc. 455, 1919 (2016).
  25. J. S. Speagle, dynesty: A dynamic nested sampling package for estimating Bayesian posteriors and evidences, Mon. Not. R. Astron. Soc. 493, 3132 (2020).
  26. H. Sun, K. L. Bouman, P. Tiede, J. J. Wang, S. Blunt, and D. Mawet, Alpha-deep probabilistic inference (α-dpi): Efficient uncertainty quantification from exoplanet astrometry to black hole feature extraction, Astrophys. J. 932, 99 (2022).
  27. M. Ruth, Exoplanet orbital characterization using simulation-based inference, Ph.D. thesis, University of Liège, Liège, Belgium, 2024.
  28. C. Chen, L. Kong, G. Li, and M. Tao, DeepTTV: Deep learning prediction of hidden exoplanet from transit timing variations, arXiv:2409.04557.
  29. R. T. Q. Chen, B. Amos, and M. Nickel, Learning neural event functions for ordinary differential equations, in International Conference on Learning Representations (ICLR, Virtual Event, Austria, 2021).
  30. X. Liu, C. Gong, and Q. Liu, Flow straight and fast: Learning to generate and transfer data with rectified flow, in The Eleventh International Conference on Learning Representations (2022).
  31. J. Wildberger, M. Dax, S. Buchholz, S. R. Green, J. H. Macke, and B. Schölkopf, Flow matching for scalable simulation-based inference, Adv. Neural Inf. Process. Syst. 36, 16837 (2023).
  32. Y. Lipman, R. T. Q. Chen, H. Ben-Hamu, M. Nickel, and M. Le, Flow matching for generative modeling, in The Eleventh International Conference on Learning Representations (ICLR, Kigali, Rwanda, 2023).
  33. G. Papamakarios and I. Murray, Fast ɛ-free inference of simulation models with Bayesian conditional density estimation, in Advances in Neural Information Processing Systems (Curran Associates, Red Hook, NY, 2016), Vol. 29, pp. 1028–1036.
  34. J.-M. Lueckmann, P. J. Goncalves, G. Bassetto, K. Öcal, M. Nonnenmacher, and J. H. Macke, Flexible statistical inference for mechanistic models of neural dynamics, arXiv:1711.01861.
  35. M. Gabrié, G. M. Rotskoff, and E. Vanden-Eijnden, Adaptive Monte Carlo augmented with normalizing flows, Proc. Natl. Acad. Sci. USA 119, e2109420119 (2022).
  36. K. W. k. Wong, M. Gabrié, and D. Foreman-Mackey, flowMC: Normalizing flow enhanced sampling package for probabilistic inference in JAX, J. Open Source Softw. 8, 5021 (2023).
  37. E. L. Nielsen, R. J. De Rosa, J. J. Wang, J. Sahlmann, P. Kalas, G. Duchêne, J. Rameau, M. S. Marley, D. Saumon, B. Macintosh, et al., The Gemini Planet Imager Exoplanet Survey: Dynamical mass of the exoplanet β Pictoris b from combined direct imaging and astrometry, Astron. J. 159, 71 (2020).
  38. See Supplemental Material at http://link.aps.org/supplemental/10.1103/ykfh-cdzk for additional details on the priors, noise models, training procedure, posterior estimates, and comparisons with neural posterior estimation methods.
  39. R. G. Belleman, J. Bédorf, and S. F. P. Zwart, High performance direct gravitational N-body simulations on graphics processing units II: An implementation in CUDA, New Astron. 13, 103 (2008).
  40. A. Lacquement, H. Beust, V. Faramaz-Gorka, and G. Duchêne, Dynamics of the β Pictoris planetary system and possibility of an additional planet, Astron. Astrophys. 694, A236 (2025).
  41. S. Lacour, J. Wang, L. Rodet, M. Nowak, J. Shangguan, H. Beust, A.-M. Lagrange, R. Abuter, A. Amorim, R. Asensio-Torres, et al., The mass of β Pictoris c from β Pictoris b orbital motion, Astron. Astrophys. 654, L2 (2021).
  42. B. Liang, C. Liu, H. Song, T. Zhao, Y. Xu, Z. Xiao, M. Liang, M. Du, W.-L. Qian, L.-e Qiang, et al., FM-MCMC: Estimating orbital parameters of direct imaging exoplanets using neural networks, GitHub repository (2026), https://github.com/aspirinliang/Estimating-Orbital-Parameters-of-Direct-Imaging-Exoplanet-Using-Neural-Network, accessed 26 May 2026.

Outline

Information

Phys. Rev. Research 8, 023341– Published 25 June, 2026
  • Received 3 December 2025
  • Accepted 27 May 2026
Crossmark - Check for updates button
Creative Commons Logo

Published by the American Physical Society under the terms of the Creative Commons Attribution 4.0 International license. Further distribution of this work must maintain attribution to the author(s) and the published article's title, journal citation, and DOI.

Published by the American Physical Society

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation