A5008296842- Particle Detector Development and Performance
- Particle physics theoretical and experimental studies
- Advancements in Semiconductor Devices and Circuit Design
- stochastic dynamics and bifurcation
- Advanced Electron Microscopy Techniques and Applications
- Diffusion and Search Dynamics
- Gaussian Processes and Bayesian Inference
- Generative Adversarial Networks and Image Synthesis
- Radiation Effects in Electronics
University of Lisbon
2022
Sapienza University of Rome
2021
Abstract Multidimensional efficiency maps are commonly used in high-energy physics experiments to mitigate the limitations generation of large samples simulated events. Binned however strongly limited by statistics. We propose a neural network approach learn ratios local densities estimate an optimal fashion efficiencies as function set parameters. Graph techniques account for high dimensional correlations between different objects event. show specific toy model how this method is applicable...
In this paper we propose a methodology for the efficient implementation of machine learning (ML)-based methods in particle-in-cell (PIC) codes, with focus on Monte Carlo or statistical extensions to PIC algorithm. The presented approach allows neural networks be developed Python environment, where advanced ML tools are readily available proficiently train and test them. Those models then efficiently deployed within highly scalable fully parallelized simulations during runtime. We demonstrate...
Multidimensional efficiency maps are commonly used in high energy physics experiments to mitigate the limitations generation of large samples simulated events. Binned multidimensional however strongly limited by statistics. We propose a neural network approach learn ratios local densities estimate an optimal fashion efficiencies as function set parameters. Graph techniques account for dimensional correlations between different objects event. show specific toy model how this method is...
This study uses neural networks to improve Monte Carlo (MC) implementations of the Bethe-Heitler process in Particle-In-Cell (PIC) codes. We provide a network that is as accurate pre-calculated tables, and requires hundred times less memory store. It trained predict pair production cross-sections for atomic numbers 1-50 photon energies between 1 MeV 10 GeV PIC code OSIRIS. first validate our approach against theoretical estimate simplified context. later prove both approaches have similar...
In this paper we propose a methodology for the efficient implementation of Machine Learning (ML)-based methods in particle-in-cell (PIC) codes, with focus on Monte-Carlo or statistical extensions to PIC algorithm. The presented approach allows neural networks be developed Python environment, where advanced ML tools are readily available proficiently train and test them. Those models then efficiently deployed within highly-scalable fully parallelized simulations during runtime. We demonstrate...