EVENTO
Strategies and techniques for deep learning on small data
Tipo de evento: Defesa de Dissertação de Mestrado
The design of models enable the interpretation of complex problems. In computer Science, such models lead to the conception of algorithms that lead to their implementation in computer systems and contribute to the prioblem solution. However, some problems are too complex to be described using an algorithmic approach. The introduction of machine learning methods aims to create models based directly on the collected data representing the observed pehnomenon. While this approach led to great advances in many different fields, data driven methods often require a substantial amount of data in order to generalize. In this work we then investigate the problem of small data for deep learning methods. We present strategies to minimize uncertainty on prediction by minimizing intra-class variation in classification tasks, constraining the solution space based on prior knowledge on the domain. Additionally, we discuss the few shot and zero shot scenarios, where we aim at traininimg robust classifiers trough a fixed kernel function in order to create a model that generalizes for classes it was not trained upon. We present experiments for each of these and evaluate their properties on distinct datasets.Para assistir essa defesa acesse: https://us02web.zoom.us/j/85927413045?pwd=RGc1aUF3NmZlL1lJSFRxTXYvSjIzQT09
Data Início: 11/09/2020 Hora: 09:00 Data Fim: 11/09/2020 Hora: 12:00
Local: LNCC - Laboratório Nacional de Computação Ciêntifica - Webinar
Aluno: Rafael Silva Pereira - - LNCC
Orientador: Fabio André Machado Porto - Laboratório Nacional de Computação Científica - LNCC
Participante Banca Examinadora: Alexis Joly - - INRIA - França Artur Ziviani - Laboratório Nacional de Computação Científica - LNCC Eduardo Bezerra da Silva - Centro Federal de Educação Tecnológica Celso Suckow da Fonseca - CEFET-RJ Fabio André Machado Porto - Laboratório Nacional de Computação Científica - LNCC
