Please use this identifier to cite or link to this item:
Title: Análisis de rendimiento del entrenamiento de una red neuronal distribuida para la clasificación de secuencias de ADN
Author: Fernández Morán, Borja
Tutor: Iserte, Sergio  
Others: Jorba, Josep  
Abstract: Neural networks for virus sequence classification can be useful for health authorities to act as quickly as possible in response to virus outbreaks or new variants of viruses. However, the large amount of input data required, as well as the complexity of the genetic sequences themselves (they may consist of tens of thousands of amino acids each), make training such a neural network a computationally expensive problem. In this context, the aim of the present work is to analyze the feasibility of a convolutional neural network for DNA sequence classification whose training is divided into several computational nodes. To perform this analysis, a convolutional neural network has been constructed that classifies genetic sequences of several different viruses. On this basis, we intend to compare a version whose training is divided into several computational nodes with respect to the traditional version trained in a single node, in order to check if there is an improvement in the total training time without affecting the rest of the network performance metrics.
Keywords: distributed computing
deep learning
DNA classification
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 17-Jan-2022
Publication license: CC BY-NC-ND  
Appears in Collections:Trabajos finales de carrera, trabajos de investigación, etc.

Files in This Item:
File Description SizeFormat 
bfernm04TFM0122memoria.pdfMemoria del TFM735,59 kBAdobe PDFThumbnail