What: phenomgrav astrophysics seminar

When: Jan 8th, 2026, 15:00-16:00 CET

Where: Building C - Room 248  (first floor)

Speaker: Dr. Mauro Pieroni, Instituto de Estructura de la Materia (IEM-CSIC), Madrid

Title: Simulation-Based Inference for Stochastic Gravitational Wave Background Reconstruction

Abstract: Simulation-Based Inference (SBI), also known as likelihood-free inference, is an alternative to traditional Monte Carlo techniques for performing Bayesian inference. SBI typically relies on machine learning methods to approximate the posterior distribution of model parameters given observed data. In recent years, SBI has been applied to a wide range of problems in physics, including gravitational wave (GW) astronomy. In this talk, I will review the basic principles of SBI and discuss its application to the detection and characterization of stochastic gravitational wave backgrounds with LISA.

What: astrophysics seminar

When: Nov 27th, 2025, 15:00-16:00 CET

Where: Building C - Room 248  (first floor)

Speaker: Dr. Annalisa Pillepich -  Max-Planck-Institut für Astronomie (MPIA)

Title: Universe(s) in a box: simulating galaxies to connect theory and observations

Abstract: Over the past decade, cosmological simulations have become a cornerstone of theoretical astrophysics — powerful “universes in a box” that connect fundamental physics and our understanding of the Universe as a whole to the galaxies and structures observed across cosmic time. These simulations combine the laws of gravity, magneto-hydrodynamics, and gas and atomic physics with models for star formation, supernova explosions, and feedback from supermassive black holes, enabling direct links between physical theory and astronomical observations. I will discuss the IllustrisTNG project, a suite of large-volume simulations that has become a benchmark for modeling the formation and evolution of galaxies, and present some of its successors, such as TNG-Cluster, which follows the growth of hundreds of massive galaxy clusters. I will highlight a few key discoveries enabled by these simulations, particularly the tight interplay between feedback from the innermost regions of galaxies and the thermodynamical, chemical, and kinematic properties of the gas in their surroundings — from the circumgalactic medium to the hot plasma of galaxy clusters. Finally, I will discuss how new techniques -- including forward modeling, machine learning, and our new project of sub-parsec resolution simulations -- are opening a path toward connecting the physics of individual stars and the interstellar medium to the formation of galaxies in the full cosmological context, as well as the use of galaxies, groups, and clusters as cosmological probes.

What: phenomgrav astrophysics seminar

When: July 7th, 2025, 15:00-16:00 CEST

Where: Building C - Room 131, ground floor

Speaker: Dr. Stefano Rinaldi - Institut für Theoretische Astrophysik, Universität Heidelberg

Title: Expect the unexpected – Inference of the black hole distribution using both parametric and non-parametric methods

Abstract: Black holes are the remnants of the most massive of stars, and characterising their population using gravitational wave (GW) observations can help us understand the astrophysical processes governing the life and death of stars. In this picture, non-parametric methods represent an excellent agnostic tool to describe our observations and guide the development of new astrophysical models; on the downside, they lack the interpretability carried by physically informed models. During this talk, I will present two approaches designed to combine the strengths of parametric and non-parametric methods: 1) the concept of augmented mixture model – a weighted superposition of astrophysical and non-parametric models capable of accounting for unforeseen features in the black hole spectrum - and 2) a procedure that first performs a non-parametric (data-driven) reconstruction of the underlying distribution, and then remaps these results onto a posterior for the parameters of an informed model.

For previous seminars, see here.