Advanced search
Start date
Betweenand

On the study of semantics in deep learning models

Grant number: 15/25739-4
Support type:Scholarships in Brazil - Master
Effective date (Start): March 01, 2016
Effective date (End): July 01, 2018
Field of knowledge:Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
Principal Investigator:João Paulo Papa
Grantee:Gustavo Henrique de Rosa
Home Institution: Faculdade de Ciências (FC). Universidade Estadual Paulista (UNESP). Campus de Bauru. Bauru , SP, Brazil
Associated scholarship(s):16/21243-7 - Learning dropout parameters for convolutional neural networks, BE.EP.MS

Abstract

Deep learning architectures have been extensively studied in the last years, mainly due to their discriminative power and effectiveness in many crucial problems in computer vision, such as face and people identification, as well as object recognition, just to name a few. However, one problem related to these models concerns with their number of parameters, which can easily reach thousands of hundreds. Another drawback is related to the need for large datasets for training purposes, as well as their high probability of overtraining, mainly because of their complex architecture. Although some recent works have proposed different solutions to alleviate this problem, such approaches still require parameters, which need to be fine-tuned, and also are skilled-dependent. This proposal aims at "learning how these techniques do learn", i.e., to learn their good points and shortcomings. Techniques such as Restricted Boltzmann Machines and Deep Belief Nets will be studied in order to obtain a better understanding of their working mechanism through meta-heuristic-based optimization. This proposal also comprises an internship at Middlesex University, United Kingdom.

Scientific publications (7)
(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)
AMORIM, WILLIAN PARAGUASSU; ROSA, GUSTAVO HENRIQUE; THOMAZELLA, ROGERIO; COGO CASTANHO, JOSE EDUARDO; LOFRANO DOTTO, FABIO ROMANO; RODRIGUES JUNIOR, OSWALDO PONS; MARANA, APARECIDO NILCEU; PAPA, JOAO PAULO. Semi-supervised learning with connectivity-driven convolutional neural networks. PATTERN RECOGNITION LETTERS, v. 128, p. 16-22, DEC 1 2019. Web of Science Citations: 0.
AMORIM, WILLIAN PARAGUASSU; TETILA, EVERTON CASTELAO; PISTORI, HEMERSON; PAPA, JOAO PAULO. Semi-supervised learning with convolutional neural networks for UAV images automatic recognition. COMPUTERS AND ELECTRONICS IN AGRICULTURE, v. 164, SEP 2019. Web of Science Citations: 0.
PASSOS, LEANDRO A.; DE SOUZA, JR., LUIS A.; MENDEL, ROBERT; EBIGBO, ALANNA; PROBST, ANDREAS; MESSMANN, HELMUT; PALM, CHRISTOPH; PAPA, JOAO PAULO. Barrett's esophagus analysis using infinity Restricted Boltzmann Machines. JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION, v. 59, p. 475-485, FEB 2019. Web of Science Citations: 0.
DE ROSA, GUSTAVO H.; PAPA, JOAO P.; YANG, XIN-S. Handling dropout probability estimation in convolution neural networks using meta-heuristics. SOFT COMPUTING, v. 22, n. 18, SI, p. 6147-6156, SEP 2018. Web of Science Citations: 0.
PEREIRA, CLAYTON R.; PEREIRA, DANILO R.; ROSA, GUSTAVO H.; ALBUQUERQUE, VICTOR H. C.; WEBER, SILKE A. T.; HOOK, CHRISTIAN; PAPA, JOAO P. Handwritten dynamics dynamics assessment through convolutional neural networks: An application to Parkinson's disease identification. ARTIFICIAL INTELLIGENCE IN MEDICINE, v. 87, p. 67-77, MAY 2018. Web of Science Citations: 15.
PAPA, JOAO PAULO; ROSA, GUSTAVO HENRIQUE; PAPA, LUCIENE PATRICI. A binary-constrained Geometric Semantic Genetic Programming for feature selection purposes. PATTERN RECOGNITION LETTERS, v. 100, p. 59-66, DEC 1 2017. Web of Science Citations: 3.
PAPA, JOAO PAULO; ROSA, GUSTAVO H.; PEREIRA, DANILLO R.; YANG, XIN-SHE. Quaternion-based Deep Belief Networks fine-tuning. APPLIED SOFT COMPUTING, v. 60, p. 328-335, NOV 2017. Web of Science Citations: 6.
Academic Publications
(References retrieved automatically from State of São Paulo Research Institutions)
ROSA, Gustavo Henrique de. Otimização Meta-Heurística para Regularização de Modelos de Aprendizado em Profundidade. 2018. 90 f. Master's Dissertation - Universidade Estadual Paulista "Júlio de Mesquita Filho" Instituto de Biociências, Letras e Ciências Exatas..

Please report errors in scientific publications list by writing to: cdi@fapesp.br.