Bachelor thesis with Prof. Lesgourgues in 2018
Recent ideas concerning non-trivial DM models
People are proposing increasingly complex Dark Matter models from the point of view of their implications on cosmological observations. Decoupled Cold Dark Matter would be order zero in complexity. Decaying, annihilating and scattering Dark Matter are more complex but are already well studied. At the next level, people are now proposing models incorporating even more refined testable features. The student will review the recent literature on this, try to understand the main features of these models, and interact with a PhD student knowing how to implement them in a simulation code in view of future research work.
Skills required: The work is essentially bibliographical and theoretical.
Skills learned: This is an occasion to learn on both cosmology and particle physics.
Main supervisor: Prof. Julien Lesgourgues (together with MSc. Deanna Hooper)
Neutrino Wars: the Last Prior
Cosmological constraints on the neutrino mass sum are currently much stronger than laboratory limits. However, these constraints are obtained through Bayesian inference and thus depend on the choice of the prior. Different authors currently find different results and there is a big controversy about that. The goal of the project is to determine the impact of different priors and to implement a better one motivated by fundamental physics.
Skills required: The code requires some programming skills, especially a good knowledge of Python!
Skills learned: This is an occasion to learn about statistical methods.
Main supervisor: Dr. Maria Archidiacono (together Prof. Julien Lesgourgues)
Relativistic Analysis of Cosmological Simulations
Recently the relativistic particle simulation gevolution and the relativistic ray-tracing code LIGER have been released. The goal of this project is to employ both codes in combination, run a few smaller simulations and perform a relativistic analysis to identify the significance of various signatures in the output data that are not present in a usual Newtonian analysis.
Skills required and learned (in decreasing importance): Data Analysis, Statistical methods, Numerical techniques, Coding
Main supervisor: Dr. Christian Fidler
A Fishy alternative
Constraints on cosmological parameters are commonly inferred using Bayesian statistics and Markov Chain Monte Carlo simulations. This can be a time consuming process, and for some complicated models it can be computationally prohibitive. Using Fisher information matrices we can quickly obtain constraints in the Gaussian approximation. The project involves improving the precision of current Fisher matrix implementations to the point where it can be used for research, and then to apply the method to interesting, but computationally expensive, cosmological models.
Skills required: Experience with Python programming
Skills learned: Statistics (Fisher information matrix, Bayesian statistics, Markov Chain Monte Carlo), Python programming, introductory cosmology
Supervisors: Prof. Julien Lesgourgues (together with MSc. Thejs Brinckmann)
Web release of a CMB graphical interface
Skills learned: the student will both learn on the underlying physics (i.e. the theory of CMB perturbations) and on the creation of interactive web pages.
Supervisors: Julien Lesgourgues