On explainability of deep neural networks

Deep Learning has attained state-of-the-art performance in the recent years, but it is still hard to determine the reasoning behind each prediction. This project will cover the latest advances on interpretability and propose a new method for pixel attribution on image classifiers.

Detalles Bibliográficos
Autor: Parafita Martínez, Álvaro
Tipo de recurso: tesis de maestría
Fecha de publicación:2018
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2117/121638
Acceso en línea:https://hdl.handle.net/2117/121638
Access Level:acceso abierto
Palabra clave:Neural networks (Computer science)
Machine learning
interpretabilitat
explicabilitat
visualització de característiques
atribució
DL
ML
CNN
interpretability
explainability
feature visualization
attribution
xarxes neuronals
xarxes neuronals convolucionals
machine learning
neural networks
convolutional neural networks
deep learning
Xarxes neuronals (Informàtica)
Aprenentatge automàtic
Àrees temàtiques de la UPC::Informàtica
Descripción
Sumario:Deep Learning has attained state-of-the-art performance in the recent years, but it is still hard to determine the reasoning behind each prediction. This project will cover the latest advances on interpretability and propose a new method for pixel attribution on image classifiers.