n o n l i n e a r b a y e s i a n i n f e r e n c e o f c
play

N o n - l i n e a r b a y e s i a n i n f e r - PowerPoint PPT Presentation

Jan 08, 2023 •17 likes •327 views

N o n - l i n e a r b a y e s i a n i n f e r e n c e o f c o s m i c f e l d s i n S D S S 3 a n d 2 M + + Guilhem Lavaux (IAP/CNRS) and Aquila Consortium Statistical Challenge for large-scale

  1. N o n - l i n e a r b a y e s i a n i n f e r e n c e o f c o s m i c f e l d s i n S D S S 3 a n d 2 M + + Guilhem Lavaux (IAP/CNRS) and Aquila Consortium Statistical Challenge for large-scale structure in the era of LSST (Oxford 2018) Aquila consortium (https://aquila-consortium.org)

  2. Outline Outline T h e s t a t i s t i c a l f r a m e w o r k T h e 2 M + + c o m p i l a t i o n ( p r e s e n t a t i o n , c l u s t e r s , v e l o c i t y f e l d s , a p p l i c a t i o n s ) S D S S 3 B O S S ( m o r e m o d e l i n g c h a l l e n g e s , d e n s i t y f e l d ) C o n c l u s i o n

  3. From theory to observations... From theory to observations... Model Observations Perfect Great but messy Complete description We do not understand the physics Full knowledge of physics Systematics not fully known Did I say perfect ? Good attempt by observers to seemingly make our life easier end up bad Various hacking to make sense of data

  4. From theory to observations... From theory to observations... Model Observations Perfect Great but messy Complete description We do not understand the physics Full knowledge of physics Systematics not fully known Did I say perfect ? Good attempt by observers to seemingly make our life easier end up bad BORG3 Still far too perfect though… (see later) Another perspective to automatically solve this problem: see Tom Charnock’s talk

  5. The BORG3 inference framework The BORG3 inference framework Forward description Observations Adjoint gradient Initial conditions Encode survey systematic effects with expansions: Jasche et al. (2010), Jasche & Wandelt (2013), Lavaux & Jasche (2016), Jasche & Lavaux (2017, 2018)

  6. BORG-PM Performance aspect BORG-PM Performance aspect 100 Hyperthreading Time (seconds) 10 70 700 Number of cores BORG-(2)LPT is ~20 times faster

  7. A p p l i c a t i o n t o 2 M + + g a l a x y c o m p i l a t i o n : D e t a i l e d d y n a m i c a l m o d e l i n g

  8. The 2M++ galaxy compilation The 2M++ galaxy compilation 0 Mpc/h SDSS Galaxy distribution 2MRS 250 Mpc/h Redshift completeness 6dF ~70 000 galaxies Lavaux & Hudson (MNRAS, 2011)

  9. Inferred density fields Inferred density fields Ensemble average density fields at z=0 Clusters Higher error Mean Void Jasche & Lavaux (2018, in prep.)

  10. Performance aspect (2): burnin Performance aspect (2): burnin Jasche & Lavaux (2018, in prep.)

  11. Initial condition powerspectrum Initial condition powerspectrum Initial conditions Post PM simulation Jasche & Lavaux (2018, in prep.)

  12. Virgo cluster Virgo cluster Virgo center ~30 Mpc/h Jasche & Lavaux; Lavaux & Jasche; Peirani, Lavaux & Jasche (2018, in prep.)

  13. Coma dynamical properties Coma dynamical properties Coma center Jasche & Lavaux; Lavaux & Jasche; Peirani, Lavaux & Jasche (2018, in prep.)

  14. Coma dynamical properties Coma dynamical properties Zoom simulation on Coma (~250 Mpart in zoom) Jasche & Lavaux; Lavaux & Jasche; Peirani, Lavaux & Jasche (2018, in prep.)

  15. Shapley concentration Shapley concentration Single realisation density Shapley center ~60 Mpc/h Lavaux & Jasche (2018, in prep.)

  16. Shapley concentration Shapley concentration Ensemble average density Shapley center ~60 Mpc/h Lavaux & Jasche (2018, in prep.)

  17. Inferred velocity fields Inferred velocity fields 800 km/s +400 km/s Outfall Higher error Infall -400 km/s 0 km/s Jasche & Lavaux (2018, in prep.)

  18. Velocity field and Hubble constant Velocity field and Hubble constant Mean error on Hubble measurement using tracers from observed large scale structures Jasche & Lavaux (2018, in prep.)

  19. A p p l i c a t i o n t o S l o a n D i g i t a l S k y S u r v e y I I I : D e e p c o s m o l o g i c a l a p p l i c a t i o n

  20. SDSS3 data SDSS3 data Panstarrs SDSS DR12 galaxy sample ~1.6 millions of galaxies SDSS

  21. Forward model becomes more complex Forward model becomes more complex Cosmic growth of structures Cosmic expansion Implemented so far for (2)LPT: Non-linear density remapping: (see Doogesh’ talk)

  22. Forward model becomes more complex Forward model becomes more complex Cosmic growth of structures Cosmic expansion L o o k b a c k t i m e Time (see Doogesh’ talk)

  23. Forward model becomes more complex Forward model becomes more complex Cosmic growth of structures Cosmic expansion L o o k b a c k t i m e Time (see Doogesh’ talk)

  24. Some systematic cleaning… Some systematic cleaning… 11 foregrounds (here only 8)… still much less than Leistedt & Peiris (2014) but improving DUST psfWidth Sky fluxes 0 ... Star densities ...

  25. Example fitted composite... Example fitted composite... 11 foregrounds (here only 8)… still much less than Leistedt & Peiris (2014) but improving DUST psfWidth Sky fluxes 0 ... Star densities ... Preliminary

  26. Inferred density of SDSS3 Inferred density of SDSS3 Ensemble density average Error estimate from ensemble variance SGC NGC Main galaxy sample limit (not included) LOWZ limit CMASS limit Preliminary

  27. Sky density Sky density Preliminary

  28. C o n c l u s i o n

  29. The Aquila consortium The Aquila consortium ● Founded in 2016 ● Gather people interested in working with each other on developing the Bayesian pipelines and run analysis on data. https://aquila-consortium.org/

  30. Conclusion: great future Conclusion: great future 2M++ SDSS CosmicFlows LSST? 2M++ SDSS CosmicFlows LSST? BORG3+ BORG3+ Predictive cosmology Cosmological measurement Predictive cosmology Cosmological measurement ● Velocity field (also VIRBIUS with F. Fuhrer) ● Cosmic expansion (see Doogesh’s talk) ● X-ray cluster emission ● Power spectrum (and governing parameters) ● Kinetic Sunyaev Zel’dovich ● Gaussianity tests of initial conditions ● Rees-Sciama ● Direct probe of dynamics ● Dark matter ?

  31. Conclusion: great future and challenges and challenges Conclusion: great future 2M++ SDSS CosmicFlows LSST? 2M++ SDSS CosmicFlows LSST? BORG3+ BORG3+ Predictive cosmology Cosmological measurement Predictive cosmology Cosmological measurement Galaxy formation: bias and likelihood Galaxy formation: bias and likelihood ● Velocity field (also VIRBIUS with F. Fuhrer) ● Cosmic expansion (see Doogesh’s talk) ● X-ray cluster emission ● Power spectrum (and governing parameters) Instrument modeling Instrument modeling ● Kinetic Sunyaev Zel’dovich ● Gaussianity tests of initial conditions ● Rees-Sciama ● Direct probe of dynamics ● Dark matter ?

Recommend

More recommend