24–28 Aug 2026
Leiden University
Europe/Zurich timezone

Searching for Cosmic Topology with Bayesian Methods using the Likelihood

Not scheduled
20m
Gorlaeus gebouw (Leiden University)

Gorlaeus gebouw

Leiden University

Einsteinweg 55, 2333 CC Leiden
Talk Methods / Statistical Inference / Machine Learning

Speaker

Linn Htat Lu (Imperial College London)

Description

The universe has a non-trivial topology if there exist non-contractible closed loops through space. If the length scale of the topology is comparable to that of the last scattering surface, then this would influence the statistics of the cosmic microwave background (CMB). Here, we discuss one method of searching for this signal—the likelihood search.
The covariance matrix describing the CMB statistics can be calculated and used to find the Gaussian likelihood of the topological model parameters given a CMB map. This includes not only the length scales and angles of the axes of a given topology, but also (since topologies are generally anisotropic) the orientation of the observer. We discuss methods to efficiently compute the likelihood and explore this parameter space.
The covariance matrices can also be used to calculate the Kullback–Leibler (KL) divergence. We can also calculate the Fisher matrix for each topology, which gives us an estimate of the error bars.
These techniques are used to show which topologies are allowed and discoverable given current Planck CMB limits and determine the extent to which future CMB experiments like LiteBIRD can improve these limits.

Other topic / keywords: Cosmic Topology

Authors

Mr Amirhossein Samandar (Case Western Reserve University) Prof. Andrew Jaffe (imperial college London) Prof. Craig Copi (Case Western Reserve University) Glenn David Starkman (Case Western Reserve University (US)) Linn Htat Lu (Imperial College London)

Co-author

Presentation materials

There are no materials yet.