Gianluca Gregori

Abstract:

The 2018 archeological survey, which was conducted at Pieve near Orzivecchi (Brescia), revealed a diverse array of archeological materials ranging from several historical periods. In particular, an altar dedicated to the indigenous god Bolgolius by Tertius Donnedo Tertulli f. was discovered. According to linguistic theory, the theonym Bolgolius probably has Celtic origins and perhaps has a connection with the god Mercurius.

Abstract:

The so-called ‘Casa del Moralista’ stands out from other houses in Pompeii on ac-count of its summer triclinium’s parietal decoration; here, we find three metrical inscriptions which are unique in the Campanian city and rare in general. One elegiac couplet is found on each wall. These texts will be evaluated from a literary point of view, but also within their immediate environmental context, to understand whether their arrangement on the three walls of the room is random, or follows a logical order.

Abstract:

Computer simulations are invaluable tools for scientific discovery. However, accurate simulations are often slow to execute, which limits their applicability to extensive parameter exploration, large-scale data analysis, and uncertainty quantification. A promising route to accelerate simulations by building fast emulators with machine learning requires large training datasets, which can be prohibitively expensive to obtain with slow simulations. Here we present a method based on neural architecture search to build accurate emulators even with a limited number of training data. The method successfully accelerates simulations by up to 2 billion times in 10 scientific cases including astrophysics, climate science, biogeochemistry, high energy density physics, fusion energy, and seismology, using the same super-architecture, algorithm, and hyperparameters. Our approach also inherently provides emulator uncertainty estimation, adding further confidence in their use. We anticipate this work will accelerate research involving expensive simulations, allow more extensive parameters exploration, and enable new, previously unfeasible computational discovery.