Skip to content

Utilizing Python and ICA, this project detects non-Gaussian signals in simulated early-universe fields to uncover insights into cosmic inflation. The tool, tested on 1D, 2D, and 3D data, combines advanced ML and computational techniques to enhance cosmological models and develop innovative data analysis methods for scientific research.

Notifications You must be signed in to change notification settings

inthaider/png_ml_project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

69 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Needle in a Haystack:

Characterizing Spatially Localized & Intermittent Primordial non-Gaussianity (PNG)

Developing an ML (Machine Learning) pipeline to disentangle PNG from simulated and observed cosmological data.

ICA-based Analysis of Primordial Non-Gaussianities (PNG) in Cosmic Data

Overview: This Python-based project, utilizing Independent Component Analysis (ICA), investigates primordial non-Gaussianities (PNG)—subtle irregularities in the universe's earliest light and matter patterns. These irregularities, which deviate from the standard Gaussian distribution expected from the early universe, hold clues about the dynamics of cosmic inflation and the universe’s rapid expansion moments after the Big Bang. By analyzing these features within simulated cosmic data, our tool, dubbed 'Cosmic-ICA' or CICA, aims to refine our understanding of the universe's infancy. Currently tested on 1D, 2D, and 3D simulated data, this approach is a precursor to applications on real observational data, paving the way for groundbreaking insights into the foundational physics of our cosmos.

Project Goals:

  • Pioneer Advanced Data Analysis: Deploy the CICA algorithm to dissect complex simulations, setting the stage for future analyses on actual cosmic observations.
  • Expand Cosmological Knowledge: Use machine learning to uncover and study non-Gaussian features, potentially revealing new aspects of the early universe’s structure.
  • Develop Robust Computational Tools: Enhance software capabilities to process and analyze high-dimensional cosmic data, with broader implications for data science and computational research in various scientific domains.

Technical Features:

  • Integration of Machine Learning and Statistics: Employs cutting-edge statistical methods and machine learning to detect and analyze non-Gaussian signals in cosmic fields, showcasing the synergy between computational intelligence and cosmology.
  • Python as a Core Technology: Utilizes Python for its robust, flexible capabilities in handling complex data analysis, ensuring scalability and adaptability of the project.
  • Iterative Development and Testing: The project is in a dynamic phase, rigorously tested on advanced simulations like 1D curvature perturbation (zeta) fields and multi-dimensional configurations created using PeakPatch, with ambitions to evolve and tackle real-world cosmic data.

Usage: TBD.

Contribute: We invite contributions to refine the algorithm, expand its functionality, or prepare it for groundbreaking applications on observational data. This project thrives on collaborative innovation, welcoming expertise from developers, data scientists, and cosmologists.

About

Utilizing Python and ICA, this project detects non-Gaussian signals in simulated early-universe fields to uncover insights into cosmic inflation. The tool, tested on 1D, 2D, and 3D data, combines advanced ML and computational techniques to enhance cosmological models and develop innovative data analysis methods for scientific research.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published