Abstract: Simulating nature and in particular processes in particle physics require expensive computations and sometimes would take much longer than scientists can afford. Here, we explore ways to a solution for this problem by investigating recent advances in generative modeling and present a study for the generation of events from a physical process with deep generative models. The simulation of physical processes requires not only the production of physical events, but to also ensure that these events occur with the correct frequencies. We investigate the feasibility of learning the event generation and the frequency of occurrence with several generative machine learning models to produce events like Monte Carlo generators. We study three processes: a simple two-body decay, the processes e(+)e(-)-> Z -> l(+)l(-) and pp -> tt<mml:mo><overbar></mml:mover> including the decay of the top quarks and a simulation of the detector response. By buffering density information of encoded Monte Carlo events given the encoder of a Variational Autoencoder we are able to construct a prior for the sampling of new events from the decoder that yields distributions that are in very good agreement with real Monte Carlo events and are generated several orders of magnitude faster. Applications of this work include generic density estimation and sampling, targeted event generation via a principal component analysis of encoded ground truth data, anomaly detection and more efficient importance sampling, e.g., for the phase space integration of matrix elements in quantum field theories. Here, the authors report buffered-density variational autoencoders for the generation of physical events. This method is computationally less expensive over other traditional methods and beyond accelerating the data generation process, it can help to steer the generation and to detect anomalies.

Abstract: In this paper, we present steady-state relativistic magnetohydrodynamic simulations that include a mass-load term to study the process of jet deceleration. The mass load mimics the injection of a proton-electron plasma from stellar winds within the host galaxy into initially pair plasma jets, with mean stellar mass-losses ranging from 10(-14) to 10(-9) M-circle dot yr(-1). The spatial jet evolution covers similar to 500 pc from jet injection in the grid at 10 pc from the jet nozzle. Our simulations use a relativistic gas equation of state and a pressure profile for the ambient medium. We compare these simulations with previous dynamical simulations of relativistic, non-magnetized jets. Our results show that toroidal magnetic fields can prevent fast jet expansion and the subsequent embedding of further stars via magnetic tension. In this sense, magnetic fields avoid a runaway deceleration process. Furthermore, when the mass load is large enough to increase the jet density and produce fast, differential jet expansion, the conversion of magnetic energy flux into kinetic energy flux (i.e. magnetic acceleration), helps to delay the deceleration process with respect to non-magnetized jets. We conclude that the typical stellar population in elliptical galaxies cannot explain jet deceleration in classical Fanaroff-Riley type I radio galaxies. However, we observe a significant change in the jet composition, thermodynamical parameters, and energy dissipation along its evolution, even for moderate values of the mass load.

Abstract: Addressing the origin of the astrophysical neutrino flux observed by IceCube is of paramount importance. Gamma-Ray Bursts (GRBs) are among the few astrophysical sources capable of achieving the required energy to contribute to such neutrino flux through p gamma interactions. In this work, ANTARFS data have been used to search for upward going muon neutrinos in spatial and temporal coincidence with 784 GRBs occurred from 2007 to 2017. For each GRB, the expected neutrino flux has been calculated in the framework of the internal shock model and the impact of the lack of knowledge on the majority of source redshifts and on other intrinsic parameters of the emission mechanism has been quantified. It is found that the model parameters that set the radial distance where shock collisions occur have the largest impact on neutrino flux expectations. In particular, the bulk Lorentz factor of the source ejecta and the minimum variability time-scale are found to contribute significantly to the GRB-neutrino flux uncertainty. For the selected sources, ANTARES data have been analysed by maximizing the discovery probability of the stacking sample through an extended maximum-likelihood strategy. Since no neutrino event passed the quality cuts set by the optimization procedure, 90 per cent confidence level upper limits (with their uncertainty) on the total expected diffuse neutrino flux have been derived, according to the model. The GRB contribution to the observed diffuse astrophysical neutrino flux around 100 TeV is constrained to be less than 10 percent.