Training multilayer perceptrons with a Gaussian Artificial Immune System

Abstract

In this paper we apply an immune-inspired approach to train Multilayer Perceptrons (MLPs) for classification problems. Our proposal, called Gaussian Artificial Immune System (GAIS), is an estimation of distribution algorithm that replaces the traditional mutation and cloning operators with a probabilistic model, more specifically a Gaussian network, representing the joint distribution of promising solutions. Subsequently, GAIS utilizes this probabilistic model for sampling new solutions. Thus, the algorithm takes into account the relationships among the variables of the problem, avoiding the disruption of already obtained high-quality partial solutions (building blocks). Besides the capability to identify and manipulate building blocks, the algorithm maintains diversity in the population, performs multimodal optimization and adjusts the size of the population automatically according to the problem. These attributes are generally absent from alternative algorithms, and all were shown to be useful attributes when optimizing the weights of MLPs, thus guiding to high-performance classifiers. GAIS was evaluated in six well-known classification problems and its performance compares favorably with that produced by contenders, such as opt-aiNet, IDEA and PSO.

Designing fuzzy rule bases with a Bayesian Artificial Immune System

“…Founded on this framework, some algorithms was developed and successfully applied to different optimization problems. For instance, the Bayesian Artificial Immune System (BAIS) [3] [4] [5] [11] and Multi-objective Bayesian Artificial Immune System (MOBAIS) [6][7] [8] for single and multiobjective optimization in discrete domains, respectively, and the Gaussian Artificial Immune System (GAIS) [9] [11] and Multi-objective Gaussian Artificial Immune System (MO-GAIS) [10] for single and multi-objective optimization in continuous domains, respectively. Besides the capability to deal with building blocks, these algorithms still preserve the aforementioned advantages of AISs.…”

Evaluating the performance of a Bayesian Artificial Immune System for designing fuzzy rule bases

“…Founded on this framework, some algorithms were developed and successfully applied to different optimization problems. For instance, the Bayesian Artificial Immune System (BAIS) [5][6][7]14] and Multi-objective Bayesian Artificial Immune Sys-tem (MOBAIS) [8][9][10] for single and multi-objective optimization in discrete domains, respectively, and the Gaussian Artificial Immune System (GAIS) [11,14] and Multi-objective Gaussian Artificial Immune System (MOGAIS) [12] for single and multi-objective optimization in continuous domains, respectively.…”

Research on Dynamic Clonal Selection Algorithm Combined with Artificial Fish School

“…Clonal selection study was first brought up by professor Bttnet [1] in 1959, and the clonal selection algorithm based on the principle of the clonal selection was mentioned up by De Castro in the university in Brazil in 1999 [2]. Hyper mutation and Receptor editing describe its main character.…”