Elucidating QGP properties at finite baryon density with Bayesian inference of RHIC Beam Energy Scan data

We present a systematic Bayesian analysis of Quark-Gluon Plasma (QGP) properties at finite baryon density using measurements of Au+Au collisions at the RHIC Beam Energy Scan program. The theoretical model simulates event-by-event (3+1)D dynamics of relativistic heavy-ion collisions with the state-of-the-art hybrid hydrodynamics and hadronic transport theory. We analyze the model's 20-dimensional posterior distributions, obtained using three Gaussian Process emulators with varying accuracy, and demonstrate the essential role of training an accurate model emulator in Bayesian analysis [1]. Our study provides robust constraints on the transport properties of the Quark-Gluon Plasma and various aspects of (3+1)- dimensional relativistic nuclear dynamics, using heavy-ion measurements from 7.7 to 200 GeV [2]. By running full model simulations with a few parameter sets sampled from the posterior distribution, we make timely predictions for $p_T$-differential observables, anisotropic flow rapidity decorrelation, and flow observables in O+O collisions with systematic theory uncertainties, which can be compared with the upcoming measurements from the STAR Collaboration. Finally, we present an exploratory analysis of iterative Bayesian inference using a non-trivial informative prior enabled by Normalization Flow techniques.