EVENTO
NAZCA: a machine learning based framework for performance prediction and configuration recommendation of multiscale numerical simulations
Tipo de evento: Seminário de Avaliação - Série A
Multiscale phenomena are observed in nature, which is increasingly attracting the attention of researchers from different areas. Simulations that intend to represent such phenomena should use computationally robust methods. The use of these methods is often limited to a small group of users who can understand the inherent complexity of each method. The overarching objective of this thesis is to help users of this type of simulations.With this objective in mind, we propose a framework called NAZCA. This framework is based on machine learning, thereby using a dataset of previous simulations to help users. We define several scenarios in which help for users of multiscale simulations is needed. Each of these scenarios is associated with a task, which will be solved with a specific machine learning technique. In this thesis, we consider the multiscale hybrid mixed (MHM) finite element method as our case study. For now, we have used NAZCA to analyze two tasks from two different scenarios involving the MHM method. In the first task, we intend to estimate the execution time of an MHM simulation. For this task, we developed a specific learning technique that explores specific knowledge about the numerical method. We show that this technique obtains smaller errors than other, state-of-the-art techniques, and a highlevel of interpretability. In the second task, we intend to recommend numerical parameters for an MHMsimulation, based on the execution time and numerical accuracy desired by the user. Initial results show a promising path in solving this task. We expect it will be possible to use this approach with other numerical methods with similar computational characteristics.Para assistir acesse: https://us02web.zoom.us/j/87126242982?pwd=NU1oT01sTnliU0dKV0lDeitJN2ZCUT09
Data Início: 29/07/2021 Hora: 14:00 Data Fim: 29/07/2021 Hora: 17:00
Local: LNCC - Laboratório Nacional de Computação Ciêntifica - Webinar
Aluno: Juan Humberto Leonardo Fábian - -
Orientador: Antônio Tadeu Azevedo Gomes - Laboratório Nacional de Computação Científica - LNCC Eduardo Soares Ogasawara - Centro Federal de Educação Tecnológica Celso Suckow da Fonseca - CEFET / RJ
Participante Banca Examinadora: Antônio Tadeu Azevedo Gomes - Laboratório Nacional de Computação Científica - LNCC Fabio André Machado Porto - Laboratório Nacional de Computação Científica - LNCC Frédéric Gerard Christian Valentin - Laboratório Nacional de Computação Científica - LNCC
Suplente Banca Examinadora: Bruno Richard Schulze - Laboratório Nacional de Computação Científica - LNCC
