Search Menu
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information
Search Scholarly Publications

Title: Dark Energy Survey Year 3 results: Simulation-based đ‘€CDM inference from weak lensing and galaxy clustering maps with deep learning: Analysis design

Journal Article · · Physical Review. D.
DOI: https://doi.org/10.1103/3sj1-1l9f · OSTI ID:3003541
ORCiD logo [1];  [1];  [2];  [3];  [4];  [1];  [5];  [6];  [7];  [5];  [8];  [6];  [9];  [10];  [11];  [12];  [13];  [14];  [5];  [15] more »;  [16];  [6];  [17];  [18];  [19];  [20];  [12];  [21];  [22];  [6];  [23];  [24];  [25];  [26];  [12];  [25];  [17];  [27];  [28];  [9];  [29];  [12];  [30];  [31];  [32]; ORCiD logo [33];  [34];  [35];  [35];  [12];  [5];  [36];  [37];  [15];  [38];  [39];  [40];  [41];  [25];  [15];  [42];  [43]; ;  [25];  [44];  [45];  [46];  [8];  [47];  [48];  [49];  [50];  [51];  [47];  [33];  [52];  [8];  [53];  [25];  [47];  [54];  [55];  [56];  [57];  [8];  [58];  [59];  [60];  [61];  [62];  [63];  [64];  [35];  [33];  [65];  [33];  [66];  [67];  [68];  [45];  [69];  [25] « less
  1. Zurich, ETH
  2. Munich U. Observ.; Unlisted, CH
  3. Unlisted, IT
  4. ETH, Zurich (main)
  5. Chicago U., Astron. Astrophys. Ctr.; Chicago U., EFI; Chicago U., KICP
  6. Barcelona, IEEC
  7. SLAC
  8. University Coll. London
  9. Princeton U., Astrophys. Sci. Dept.
  10. Wisconsin U., Madison
  11. Argonne (main)
  12. Pennsylvania U., Dept. Math.
  13. Carnegie Mellon U.
  14. IAC, La Laguna; LIneA, Rio de Janeiro; Laguna U., Tenerife
  15. Duke U.
  16. NASA, Goddard
  17. KIPAC, Menlo Park
  18. LBL, Berkeley
  19. Chicago U., Astron. Astrophys. Ctr.; Fermilab; Chicago U., EFI; Chicago U., KICP
  20. Pennsylvania U., Dept. Math.; LPSC, Grenoble
  21. Waterloo U.
  22. Caltech
  23. Munich U. Observ.
  24. Cardiff U.
  25. Fermilab
  26. Caltech; Caltech, JPL
  27. Cambridge U., DAMTP
  28. Princeton U., Astrophys. Sci. Dept.; KIPAC, Menlo Park; SLAC; Munich U. Observ.
  29. Campinas State U.
  30. Madrid, CIEMAT; Ruhr U., Bochum
  31. Chicago U., Astron. Astrophys. Ctr.; Nordita; Royal Inst. Tech., Stockholm
  32. INFN, Genoa
  33. Madrid, CIEMAT
  34. Jodrell Bank
  35. KIPAC, Menlo Park; SLAC
  36. Brookhaven
  37. SUNY, Stony Brook
  38. IRAP, Toulouse
  39. ORIGINS, Garching; Garching, Max Planck Inst., MPE; Munich U.
  40. BCCP, Berkeley; UC, Berkeley; LBL, Berkeley
  41. Stanford U.; KIPAC, Menlo Park; SLAC
  42. Cerro-Tololo InterAmerican Obs.
  43. Edinburgh U., Inst. Astron.
  44. Zurich U.
  45. Portsmouth U., ICG
  46. Northeastern U.
  47. Queensland U.
  48. Barcelona, IFAE
  49. Unlisted, US
  50. LIneA, Rio de Janeiro
  51. Hamburg Observ.
  52. Indian Inst. Tech., Hyderabad
  53. Madrid, IFT
  54. UC, Santa Cruz
  55. Ohio State U., CCAPP; Ohio State U.
  56. Harvard-Smithsonian Ctr. Astrophys.
  57. Australian Astron. Observ.; Lowell Observ.
  58. Caltech, JPL
  59. Texas A-M
  60. LPSC, Grenoble
  61. Illinois U., Urbana (main); Illinois U., Urbana, Astron. Dept.
  62. ICREA, Barcelona; Barcelona, IFAE
  63. Cincinnati U.; Perimeter Inst. Theor. Phys.
  64. Rio de Janeiro Observ.
  65. Madrid, CIEMAT; Zurich U.
  66. Lancaster U.
  67. ORNL, Oak Ridge (main)
  68. Illinois U., Urbana (main)
  69. Chicago U., Astron. Astrophys. Ctr.
+ Show Author Affiliations

Data-driven approaches using deep learning are emerging as powerful techniques to extract non-Gaussian information from cosmological large-scale structure. Here, this work presents the first simulation-based inference (SBI) pipeline that combines weak lensing and galaxy clustering maps in a realistic Dark Energy Survey Year 3 (DES Y3) configuration and serves as preparation for a forthcoming analysis of the survey data. We develop a scalable forward model based on the CosmoGridV1 suite of N-body simulations to generate over one million self-consistent mock realizations of DES Y3 at the map level. Leveraging this large dataset, we train deep graph convolutional neural networks on the full survey footprint in spherical geometry to learn low-dimensional features that approximately maximize mutual information with target parameters. These learned compressions enable neural density estimation of the implicit likelihood via normalizing flows in a ten-dimensional parameter space spanning cosmological đ‘€CDM, intrinsic alignment, and linear galaxy bias parameters, while marginalizing over baryonic, photometric redshift, and shear bias nuisances. To ensure robustness, we extensively validate our inference pipeline using synthetic observations derived from both systematic contaminations in our forward model and independent Buzzard galaxy catalogs. Our forecasts yield significant improvements in cosmological parameter constraints, achieving 2−3× higher figures of merit in the đ›ș𝑚 − 𝑆8 plane relative to our implementation of baseline two-point statistics and effectively breaking parameter degeneracies through probe combination. These results demonstrate the potential of SBI analyses powered by deep learning for upcoming Stage-IV wide-field imaging surveys.

Research Organization:
Fermi National Accelerator Laboratory (FNAL), Batavia, IL (United States)
Sponsoring Organization:
USDOE Office of Science (SC), High Energy Physics (HEP)
Grant/Contract Number:
89243024CSC000002
OSTI ID:
3003541
Report Number(s):
DES--2025-0952; FERMILAB-PUB--25-0749-PPD; oai:inspirehep.net:3080641; arXiv:2511.04681
Journal Information:
Physical Review. D., Journal Name: Physical Review. D.; ISSN 2470-0010; ISSN 2470-0029
Publisher:
American Physical Society (APS)Copyright Statement
Country of Publication:
United States
Language:
English

Similar Records

Related Subjects