1 of 20

CMB-INFLATE:

WP3 Mid Term Status

Carlo Baccigalupi, SISSA

2 of 20

Network Structure

3 of 20

WP3: Innovative Astrophysical Foreground Separation Approaches and Modeling

​

  • Lead: International School for Advanced Studies (SISSA)
  • Participating organisations:
    • Max Planck Institute for Astrophysics in Garching (MPA)
    • University of Nagoia (UN)
    • University of Oslo (UO)
    • University of Paris Cité (UPC)
    • University of Tokio (UT)

4 of 20

WP3: the Science Case

Galactic Foregrounds, Planck 2018, IV

5 of 20

WP3: the Science Case

Low Frequency Foregrounds, Carretti et al. 2020, Krchmalnicoff et al. 2018

6 of 20

WP3: Objectives

  • New and publicly available
    • sky models
    • innovative component separation algorithms, based on
      • Minimum Variance
      • Parametric Fitting
  • Robust and Accurate Estimation of r

7 of 20

WP3: Tasks

  • Task 3.1: Simulation of foreground polarized emissions, particularly synchrotron and dust, based on 3-dimensional galactic magnetic fields, as well low- and high-frequency observations. Leading organisation is SISSA, participating organisations are: UPC, UO, MPA
  • Task 3.2: Development, testing and validation of Minimum Variance Component Separation algorithms. We will particularly focus on the numerical performance of these methods, and test them in the presence of spatially varying foregrounds emissions. Leading organisation is MPA. Participating organisations are: MPA, UN, UT
  • Task 3.3: Development, testing and validation of Hybrid Component Separation algorithms. These relatively new approaches will need to be characterized at the level of other methods (study of residuals properties, numerical performance, impact of priors). Leading organisation is UPC, participating organisations are: SISSA, UT

8 of 20

WP3: Tasks

  • Task 3.4: Development, testing and validation of Parametric Fitting Component Separation algorithms. Teams of this project have acquired a lot of expertise on these methods. One of the limitations is the parametrization of foreground emission laws, as well as the fine tuning of free parameters. We will build new algorithms deriving the optimal number of degrees of freedom, as learnt from the data themselves. Leading organisation is UO, participating organisations are: SISSA, UP, UT
  • Task 3.5: Thorough comparison of the above cleaning methods, under various sky complexities (Task 3.1) and in the presence of possible instrumental systematic effects. Such comparison will exploit a common simulated data set, and will take the form of a data challenge. We will perform a systematic comparison of the above algorithms, by characterizing the pros and cons of each of them. Combining the developed tools may allow us to optimally clean galactic foregrounds and exploit the up-coming CMB data sets. Leading organisation is SISSA, participating organisations are MPA, UN, UO, UPC, UT, Todai, UP.
  • Task 3.6: Inclusion of a criterion for a Bi-Spectrum Component Separation methods in order to account for the non-Gaussianity of foregrounds. Leading organisation UP. Participating organisations are UN, UPC, UT.

9 of 20

WP3: Deliverables

  • D3.1 (T+46M): Public Sky Simulations
  • D3.2 (T+30M): Minimum Variance Component Separation Methods
  • D3.3 (T+30M): Parametric Fitting Component Separation Methods
  • D3.4 (T+30M): Hybrid Component Separation Methods
  • D3.5 (T+44M): Inclusion of Non-Gaussianity in Component Separation

10 of 20

WP3: Deliverables

  • D3.1 (T+46M): Public Sky Simulations
  • D3.2 (T+30M): Minimum Variance Component Separation Methods
  • D3.3 (T+30M): Parametric Fitting Component Separation Methods
  • D3.4 (T+30M): Hybrid Component Separation Methods
  • D3.5 (T+44M): Inclusion of Non-Gaussianity in Component Separation

11 of 20

WP3: Deliverables

  • D3.1 (T+46M): Public Sky Simulations
    • First Version is Done, more in progress
  • D3.2 (T+30M): Minimum Variance Component Separation Methods
    • Published New Results and Implementation, more in progress
  • D3.3 (T+30M): Parametric Fitting Component Separation Methods
    • Published Results for the LiteBIRD Collaboration ì, more in progress
  • D3.4 (T+30M): Hybrid Component Separation Methods
    • In Progress
  • D3.5 (T+44M): Inclusion of Non-Gaussianity in Component Separation
    • In Progress

12 of 20

Public Sky Simulations

13 of 20

Public Sky Simulations: Towards New Analyses

2.3, 5, 11, 13, 17, 19, 30, 40, 43, 90 GHz

radio

QUIJOTE

LSPE-STRIP

Synchrotron scaling

14 of 20

Analysis of Data and Sky Simulations

Vacher et al., 2023

15 of 20

Minimum Variance

Carones et al. for the LiteBIRD Collaboration, 2023

16 of 20

Parametric Fitting

LiteBIRD Collaboration, 2023

17 of 20

Parametric Fitting

LiteBIRD Collaboration, 2023

18 of 20

Minimum Variance & Parametric Fitting

Fuskeland et al. 2023 for the LiteBIRD Collaboration

19 of 20

WP3 Travels & Project, Mid-Term Meeting

  • Anto Lonappan (SISSA) to UT, 2022
    • T3.2, 3.4. 3.5
    • D3.2, 3.3, 3.5
  • MPA Group to UT, 2022
    • T3.2, 3.3, 3.4
    • D3.2, 3.3, 3.4
  • Carlo Baccigalupi (SISSA) to UCB, 2022, 2023
    • T3.1, 3.4
    • D3.1, 3.3
  • SISSA Group to UT, 2023
    • T3.1, 3.2, 3.4, 3.5
    • D3.1, 3.2, 3.3
  • …

20 of 20

WP3 Session

  • Tuesday, September 26th, 2023: 9am - 10:15am CEST
  • 09:00 - 09:45: Algorithms
    • 09:00 - 09:15: Sky Model, by Nicoletta Krachmalnicoff, Giuseppe Puglisi
    • 09:15 - 09:30: Minimum Variance, by Alessandro Carones
    • 09:30 - 09:45: Parametric Fitting, by Arianna Rizzieri, Josquin Errard
  • 09:45 - 10:15: Examples of Exploitations
    • 09:45 - 10:00: Gain Calibration and Foreground Cleaning
    • 10:00 - 10:15: Foregrounds and Lensing Extraction
  • …