Deep learning: ensinando a aprender
Carregando...
Arquivos
Data
Título da Revista
ISSN da Revista
Título de Volume
Editor
275
Resumo
Aprendizado profundo atualmente é considerado a principal
técnica de aprendizado de máquina na solução de problemas,
particularmente em problemas de classificação. Contudo, há
ainda muita dificuldade no entendimento e utilização dessas
técnicas. Nesse trabalho pretende-se apresentar as Redes
Neurais Profundas e as Redes Convolucionais. Como resultado
desta pesquisa, busca-se um entendimento mais sistematizado
sobre o funcionamento dessas técnicas. Pode-se concluir que a
partir das modernas técnicas de aprendizado de máquina
existentes, os problemas envolvendo variáveis mais difíceis de
serem descritas para a máquina como, por exemplo, o instinto
humano, estão ao alcance com as infinitas possibilidades de
aprendizado, como se a própria máquina fosse uma criança
dando seus primeiros passos
Deep learning is currently considered the main machine learning technique in solving problems, particularly in classification problems. However, there is still much difficulty in understanding and using these methods. In this work we intend to present the Deep Neural Networks and the Convolutional Networks. The main focus will be on understanding how these methods work. One may conclude that, based on state-of-the-art machine learning techniques, it’s possible to solve problems containing a more complex logic for computers such as the human instinct through infinite learning possibilities. Just like if the machine itself was a child giving its first steps.
Deep learning is currently considered the main machine learning technique in solving problems, particularly in classification problems. However, there is still much difficulty in understanding and using these methods. In this work we intend to present the Deep Neural Networks and the Convolutional Networks. The main focus will be on understanding how these methods work. One may conclude that, based on state-of-the-art machine learning techniques, it’s possible to solve problems containing a more complex logic for computers such as the human instinct through infinite learning possibilities. Just like if the machine itself was a child giving its first steps.
Descrição
Palavras-chave
Citação
HOSAKI, Gabriel Yuri; RIBEIRO, Douglas Francisco. Deep learning: ensinando a aprender. RGE - Revista de Gestão e Estratégia. Assis, v.1, n.3, p. 36-50, 2021. Disponível em: https://storage.googleapis.com/production-hostgator-brasil-v1-0-3/703/940703/o9vuz1RL/4dcfd66d9bbf497ab5b0067a65bc1fc1?fileName=RGE%20-%20Vol.1%20-%20N%C2%BA3%20-%202021.pdf. Acesso em: 05 de maio de 2021.