Multivariate Analysis of the Data Temporal Structure for Blind Source Separation in the Context of Nonlinear Mixtures and of Multiple Datasets

• Grant number: 17/11488-5
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: August 01, 2017
• End date: January 28, 2018
• Field of knowledge: Engineering - Electrical Engineering - Telecommunications
• Principal Investigator: Aline de Oliveira Neves Panazio
• Grantee: Denis Gustavo Fantinato
• Host Institution: Centro de Engenharia, Modelagem e Ciências Sociais Aplicadas (CECS). Universidade Federal do ABC (UFABC). Ministério da Educação (Brasil). Santo André , SP, Brazil

Abstract

In the signal processing area, the Blind Source Separation (BSS) problem occupies a prominent position in view of its versatility and its wide range of practical applications. In light of the increasing complexity and the significant growth in the amount of data, the BSS problem in the context of nonlinear mixtures and of multiple datasets form current and challenging research topics. In the first case, the nonlinear approach still lacks for a general separating structure and, even on constrained models, there are a limited set of functions that can be compensated. In the second case, the multiple datasets are capable of bringing information diversity; however, the study involving complex and nonlinear data is still incipient. In both cases, strong evidences indicate that the information encompassed in the data temporal structure is crucial for the development of these research topics.These perspectives motivate us to study a new blind source separation paradigm from a multivariate analysis (MVA) standpoint on the data temporal structure, in the context of both nonlinear mixtures and multiple datasets. In fact, by means of the powerful MVA framework, it is possible to exploit the data temporal information in a more effective manner. This will contribute to the proposition of more robust and adequate separating criteria in the two mentioned contexts. In addition, the propositions may be analyzed in complex practical scenarios, such as in brain-computer interface with multiple subjects.

Scientific publications

(5)

(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)

FANTINATO, D. G.; SILVA, D. G.; ATTUX, R.; NEVES, A.. Multivariate Shannon's entropy for adaptive IIR filtering via kernel density estimators. ELECTRONICS LETTERS, v. 55, n. 15, p. 859+, JUL 25 2019. (13/14185-2, 17/11488-5)

FERNANDEZ, STEPHANIE A.; FANTINATO, DENIS G.; MONTALVAO, JUGURTA; ATTUX, ROMIS; SILVA, DANIEL G.; IEEE. Immune-Inspired Optimization with Autocorrentropy Function for Blind Inversion of Wiener Systems. 2018 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC), v. N/A, p. 7-pg., 2018-01-01. (17/11488-5)

FANTINATO, DENIS G.; DUARTE, LEONARDO T.; DEVILLE, YANNICK; ATTUX, ROMIS; JUTTEN, CHRISTIAN; NEVES, ALINE. A second-order statistics method for blind source separation in post-nonlinear mixtures. Signal Processing, v. 155, p. 63-72, FEB 2019. (17/11488-5, 15/23424-6)

FANTINATO, DENIS G.; NEVES, ALINE; SILVA, DANIEL G.; ATTUX, ROMIS; UEDA, N; WATANABE, S; MATSUI, T; CHIEN, JT; LARSEN, J. BLIND CHANNEL EQUALIZATION OF ENCODED DATA OVER GALOIS FIELDS. 2017 IEEE 27TH INTERNATIONAL WORKSHOP ON MACHINE LEARNING FOR SIGNAL PROCESSING, v. N/A, p. 6-pg., 2017-01-01. (17/11488-5, 13/14185-2)

FANTINATO, DENIS G.; DUARTE, LEONARDO T.; DEVILLE, YANNICK; JUTTEN, CHRISTIAN; ATTUX, ROMIS; NEVES, ALINE; DEVILLE, Y; GANNOT, S; MASON, R; PLUMBLEY, MD; et al. Using Taylor Series Expansions and Second-Order Statistics for Blind Source Separation in Post-Nonlinear Mixtures. LATENT VARIABLE ANALYSIS AND SIGNAL SEPARATION (LVA/ICA 2018), v. 10891, p. 11-pg., 2018-01-01. (17/11488-5)