André Ricardo Gonçalves

Multi-task Sparse Structure Learning

Multi-task learning (MTL) aims to improve generalization performance by learning multiple related tasks simultaneously. While sometimes the underlying task relationship structure is known, often the structure needs to be estimated from data at hand. In this paper, we present a novel family of models for MTL, applicable to regression and classification problems, capable of learning the structure of task relationships. In particular, we consider a joint estimation problem of the task relationship structure and the individual task parameters, which is solved using alternating minimization. The task relationship structure learning component builds on recent advances in structure learning of Gaussian graphical models based on sparse estimators of the precision (inverse covariance) matrix. We illustrate the effectiveness of the proposed model on a variety of synthetic and benchmark datasets for regression and classification. We also consider the problem of combining climate model outputs for better projections of future climate, with focus on temperature in South America, and show that the proposed model outperforms several existing methods for the problem.

Online learning in estimation of distribution algorithms for dynamic environments

In this paper, we propose an estimation of distribution algorithm based on an inexpensive Gaussian mixture model with online learning, which will be employed in dynamic optimization. Here, the mixture model stores a vector of sufficient statistics of the best solutions, which is subsequently used to obtain the parameters of the Gaussian components. This approach is able to incorporate into the current mixture model potentially relevant information of the previous and current iterations. The online nature of the proposal is desirable in the context of dynamic optimization, where prompt reaction to new scenarios should be promoted. To analyze the performance of our proposal, a set of dynamic optimization problems in continuous domains was considered with distinct levels of complexity, and the obtained results were compared to the results produced by other existing algorithms in the dynamic optimization literature.

Modeling Alzheimer's disease cognitive scores using multi-task sparse group lasso

Alzheimers disease (AD) is a severe neurodegenerative disorder characterized by loss of memory and reduction in cognitive functions due to progressive degeneration of neurons and their connections, eventually leading to death. In this paper, we consider the problem of simultaneously predicting several different cognitive scores associated with categorizing subjects as normal, mild cognitive impairment (MCI), or Alzheimers disease (AD) in a multi-task learning framework using features extracted from brain images obtained from ADNI (Alzheimers Disease Neuroimaging Initiative). To solve the problem, we present a multi-task sparse group lasso (MT-SGL) framework, which estimates sparse features coupled across tasks, and can work with loss functions associated with any Generalized Linear Models. Through comparisons with a variety of baseline models using multiple evaluation metrics, we illustrate the promising predictive performance of MT-SGL on ADNI along with its ability to identify brain regions more likely to help the characterization Alzheimers disease progression.

A multiobjective analysis of adaptive clustering algorithms for the definition of RBF neural network centers in regression problems

A variety of clustering algorithms have been applied to determine the internal structure of Radial Basis Function Neural Networks (RBFNNs). k-means algorithm is one of the most common choice for this task, although, like many other clustering algorithms, it needs to receive the number of prototypes a priori. This is a nontrivial procedure, mainly for real-world applications. An alternative is to use algorithms that automatically determine the number of prototypes. In this paper, we performed a multiobjective analysis involving three of these algorithms, which are: Adaptive Radius Immune Algorithm (ARIA), Affinity Propagation (AP), and Growing Neural Gas (GNG). For each one, the parameters that most influence the resulting number of prototypes composed the decision space, while the RBFNN RMSE and the number of prototypes formed the objective space. The experiments found that ARIA solutions achieved the best results for the multiobjective metrics adopted in this paper.

A multi-Gaussian component EDA with restarting applied to direction of arrival tracking

This paper analyzes the application of a multi-population Gaussian-based estimation of distribution algorithm equipped with a restarting strategy and mutation, named MGcEDA, to the problem of estimating the Direction of Arrival (DOA) of time-varying plane waves impinging on a uniform linear array of sensors. This problem requires the minimization of a dynamic cost function which is non-linear, non-quadratic, multimodal and variant with respect to the signal-to-noise ratio. Experiments showed that MGcEDA was able to quickly respond to changes in the source features in scenarios with different levels of noise and number of signals. Moreover, MGcEDA outperforms a previously proposed approach in all considered experiments in terms of well known performance measures.

The influence of supervised clustering for RBFNN centers definition: a comparative study

Several clustering algorithms have been considered to determine the centers and dispersions of the hidden layer neurons of Radial Basis Function Neural Networks (RBFNNs) when applied both to regression and classification tasks. Most of the proposed approaches use unsupervised clustering techniques. However, for data classification, by performing supervised clustering it is expected that the obtained clusters represent meaningful aspects of the dataset. We therefore compared the original versions of k-means, Neural-Gas (NG) and Adaptive Radius Immune Algorithm (ARIA) along with their variants that use labeled information. The first two had already supervised versions in the literature, and we extended ARIA toward a supervised version. Artificial and real-world datasets were considered in our experiments and the results showed that supervised clustering is better indicated in problems with unbalanced and overlapping classes, and also when the number of input features is high.

Classification of Poincaré plots for temporal series of heart rate variability by using machine learning techniques

This article presents two classifiers based on machine learning methods, aiming to detect physiologic anomalies considering Poincare´ plots of heart rate variability. It was developed a preprocessing procedure to encoding the plots, based on the Cellular Features Extraction Method. Simulation of different classifiers, artificial neural networks and support vector machine, has been performed and the performance achieved was about 94%. The study shows attractive, once can be extended for other kind of graphics that represents patterns known in the health field.