A5048434210- Neural Networks and Applications
- Evolutionary Algorithms and Applications
- Metaheuristic Optimization Algorithms Research
- Advanced Multi-Objective Optimization Algorithms
- VLSI and FPGA Design Techniques
- Blind Source Separation Techniques
- Interconnection Networks and Systems
- Parallel Computing and Optimization Techniques
- VLSI and Analog Circuit Testing
- Fuzzy Logic and Control Systems
- EEG and Brain-Computer Interfaces
- Embedded Systems Design Techniques
- Advanced Memory and Neural Computing
- Distributed and Parallel Computing Systems
- Advanced Data Storage Technologies
- Software-Defined Networks and 5G
- Network Security and Intrusion Detection
- Neural dynamics and brain function
- Fault Detection and Control Systems
- Machine Learning and Data Classification
- Network Packet Processing and Optimization
- Integrated Circuits and Semiconductor Failure Analysis
- Cloud Computing and Resource Management
- Protein Structure and Dynamics
- Anomaly Detection Techniques and Applications
i2CAT
2024
Tecnológico Nacional de México
2023
Universidad de Granada
2013-2022
Fulbright University Vietnam
2022
Literaturarchiv
2019
European Research Council
2019
Elabora Consultoria (Brazil)
2019
Departamento de Educación
2016
Andalusian Health Service
2016
Dirección de Investigación y Desarrollo
2014
This paper presents a multiobjective evolutionary algorithm to optimize radial basis function neural networks (RBFNNs) in order approach target functions from set of input-output pairs. The procedure allows the application heuristics improve solution problem at hand by including some new genetic operators process. These are based on two well-known matrix transformations: singular value decomposition (SVD) and orthogonal least squares (OLS), which have been used define mutation that produce...
In the synthesis of a fuzzy system two steps are generally employed: identification structure and optimization parameters defining it. The paper presents methodology to automatically perform these in conjunction using three-phase approach construct from numerical data. Phase 1 outlines membership functions rules for specific structure, starting very simple initial topology. 2 decides new more suitable topology with information received previous step; it determines which variable number sets...
To date, clustering techniques have always been oriented to solve classification and pattern recognition problems. However, some authors applied them unchanged construct initial models for function approximators. Nevertheless, approximation problems present quite different objectives. Therefore it is necessary design new algorithms specialized in the problem of approximation. This paper presents a technique, specially designed function. problems, which improves performance approximator...
Electroencephalography (EEG) datasets are often small and high dimensional, owing to cumbersome recording processes. In these conditions, powerful machine learning techniques essential deal with the large amount of information overcome curse dimensionality. Artificial Neural Networks (ANNs) have achieved promising performance in EEG-based Brain-Computer Interface (BCI) applications, but they involve computationally intensive training algorithms hyperparameter optimization methods. Thus, an...
Diagnosis of learning difficulties is a challenging goal. There are huge number factors involved in the evaluation procedure that present high variance among population with same difficulty. usually performed by scoring subjects according to results obtained different neuropsychological (performance-based) tests specifically designed this end. One most frequent disorders developmental dyslexia (DD), specific difficulty acquisition reading skills not related mental age or inadequate...
In this paper, a systematic design is proposed to determine fuzzy system structure and learning its parameters, from set of given training examples. particular, two fundamental problems concerning modeling are addressed: 1) rule parameter optimization 2) the identification (i.e., number membership functions rules). A four-step approach build automatically presented: Step 1 directly obtains optimum rules for function configuration. 2 optimizes allocation conclusion rules, in order achieve...
This paper deals with the use of parallel processing for multi-objective optimization in applications which objective functions, restrictions, and hence also solutions can change over time. These dynamic problems appear quite different real-world relevant socio-economic impact. The procedure this is presented based on PSFGA, a evolutionary optimization. It uses master process that distributes population among processors system (that evolve their corresponding according to an island model),...
OpenFlow switching enables flexible management of enterprise network switches and experiments on regular traffic. We present in this paper a complementary design to OpenFlow's existing reference designs. apply processor based acceleration cards perform switching. describe the options report our experiment results that show 20% reduction packet delay comparable forwarding throughput compared conventional
Brain-computer interfacing (BCI) applications based on the classification of electroencephalographic (EEG) signals require solving high-dimensional pattern problems with such a relatively small number training patterns that curse dimensionality usually arise. Multiresolution analysis (MRA) has useful properties for signal in both temporal and spectral analysis, been broadly used BCI field. However, MRA increases input data. Therefore, some approaches to feature selection or reduction should...
Electroencephalography (EEG) signal classification is a challenging task due to the low signal-to-noise ratio and usual presence of artifacts from different sources. Different techniques, which are usually based on predefined set features extracted EEG band power distribution profile, have been previously proposed. However, still remains challenge, depending experimental conditions responses be captured. In this context, use deep neural networks offers new opportunities improve performance...
An analysis of the influence weight and input perturbations in a multilayer perceptron (MLP) is made this article. Quantitative measurements fault tolerance, noise immunity, generalization ability are provided. From expressions obtained, it possible to justify some previously reported conjectures experimentally obtained results (e.g., magnitudes, relation between training with ability, tolerance ability). The introduced here explicitly related mean squared error degradation presence...