Field-based physical inference from peculiar velocity tracers

We present a Bayesian hierarchical modelling approach to reconstruct the initial cosmic matter density field constrained by peculiar velocity observations. As our features model for gravitational evolution of dark connect conditions late-time observations, it reconstructs final and fields as natural byproducts. implement this field-based physical inference adapting Origin Reconstruction from Galaxies (BORG) algorithm, which explores high-dimensional posterior through use Hamiltonian Monte Carlo sampling. test self-consistency method using random sets mock tracers, assess its accuracy in more complex scenario where tracers are non-linearly evolved haloes. find that framework self-consistently infers conditions, fields, shows some robustness mis-specification. compared state-of-the-art Gaussian fields/Wiener filtering, produces accurate reconstructions. It also allows us constrain complementing aspect previous approaches based on other cosmological observables.