## Cosmic structure formation in the non-linear regime

At small scales, the three-dimensional large-scale structure contains a wealth of cosmological information which cannot be trivially extracted due to the *non-linear dynamical evolution of the density field*. In order to optimally extract information from upcoming, but also existing, surveys, I am interested in the description of structure formation in this regime.

In 2013, I proposed a fast method to *improve the performance of Lagrangian perturbation theory in the mildly non-linear regime*, based on a remapping of the approximately-evolved density field, using information extracted from *N*-body simulations (Leclercq *et al*. 2013).

I am also the author of *Simbelmynë* [SEEM-bale-mü-nay], a flexible cosmological code for forward-modelling of the large-scale structure. Simbelmynë features an efficient implementation the temporal and spatial *tCOLA* (COmoving Lagrangian Acceleration, Tassev *et al*. 2013) and *sCOLA* (Tassev *et al*. 2015) schemes for very efficient dark matter simulations. It also includes a variety of tools to create synthetic observations of galaxies, and to analyse the dynamics of the cosmic web (phase-space field estimators, cosmic web classifiers).

*left panel*) using a non-linear gravitational model. The galaxy density field (

*middle panel*) takes into account galaxy bias and redshift-space distortions. From this, observations are simulated (

*right panel*) using the survey geometry and a prescription for selection effects and missing observations.

As an end-to-end generative process for galaxy surveys data, the Simbelmynë model can be conveniently written in the form of a Bayesian Hierarchical Model, and it is particularly suitable for use with likelihood-free inference methods.