Palestrante
Gustavo Gil da Silveira
(Universidade Federal do Rio Grande do Sul (UFRGS))
Descrição
We employ machine learning techniques to investigate novel approaches for particle simulation and identification. In one front we study the possibility of using Deep Neural Networks for jet identification in the L1 trigger at HL-LHC. We perform a survey of architectures (MLP, CNN and Graph Networks) and benchmark their performance and resource consumption on FPGAs. We use the HLS4ML jet dataset to compare the results obtained in this study to previous literature on Fast Machine Learning applications on FPGAs. We also investigate the use of generative adversarial networks (GAN) for fast simulation of particle showers in electromagnetic calorimeters, which is used to reproduce energy deposition and shower shapes in multi-layer setups.
| Key Words | machine learning, trigger, deep neural networks, generative adversarial networks, electromagnetic calorimeter, FPGA |
|---|
Autores primários
Gustavo Gil da Silveira
(Universidade Federal do Rio Grande do Sul (UFRGS))
Prof.
André Sznajder
(Universidade do Estado do Rio de Janeiro)
Co-autores
Sr.
Eduardo Gressler Brock
(Universidade Federal do Rio Grande do Sul (UFRGS))
Sr.
Vitor Dos Santos Sousa
(Universidade do Estado do Rio de Janeiro)
