Dynamics of the evolution of learning algorithms by selection

IEEE Transactions on Evolutionary Computation, 2002

In genetic programming (GP), learning problems can be classified broadly into two types: those using data sets, as in supervised learning, and those using an environment as a source of feedback. Recently, an increasing amount of research has concentrated on the robustness or generalization ability of the programs evolved using GP. While some of the researchers report on the brittleness of the solutions evolved, others proposed methods of promoting robustness/generalization. It is important that these methods are not ad hoc and are applicable to other experimental setups. In this paper, learning concepts from traditional machine learning and a brief review of research on generalization in GP are presented. The paper also identifies problems with brittleness of solutions produced by GP and suggests a method for promoting robustness/generalization of the solutions in simulating learning behaviors using GP. . Previously, he was a Lecturer in knowledge management at the University of Surrey, Guildford, U.K. His research aims to apply evolutionary learning models to simulate behaviors of adaptive virtual business organizations using complexity and evolutionary artificial intelligence techniques (i.e., genetic algorithms and programming). In particular, he is interested in complexity, evolutionary data mining, and adaptive knowledge management, i.e., modeling self-organization of knowledge for web-based environments and e-business.

1999

In this paper, the absolute benefit, a measure of improvement in the fitness space, is derived from the viewpoint of fitness distribution and fitness trajectory analysis. It is used for online operator-adaptation, where the optimization of density estimation models serves as an example. A new information theory based measure is proposed to judge the accuracy of the evolved models. Further, the absolute benefit is applied to offline analysis of new gradient based operators used for coefficient adaptation in genetic programming. An efficient method to calculate the gradient information is presented.

IEEE Transactions on Evolutionary Computation, 2007

Some of the most influential factors in the quality of the solutions found by an evolutionary algorithm (EA) are a correct coding of the search space and an appropriate evaluation function of the potential solutions. EAs are often used to learn decision rules from datasets, which are encoded as individuals in the genetic population. In this paper, the coding of the search space for the obtaining of those decision rules is approached, i.e., the representation of the individuals of the genetic population and also the design of specific genetic operators. Our approach, called "natural coding," uses one gene per feature in the dataset (continuous or discrete). The examples from the datasets are also encoded into the search space, where the genetic population evolves, and therefore the evaluation process is improved substantially. Genetic operators for the natural coding are formally defined as algebraic expressions. Experiments with several datasets from the University of California at Irvine (UCI) machine learning repository show that as the genetic operators are better guided through the search space, the number of rules decreases considerably while maintaining the accuracy, similar to that of hybrid coding, which joins the well-known binary and real representations to encode discrete and continuous attributes, respectively. The computational cost associated with the natural coding is also reduced with regard to the hybrid representation. Our algorithm, HIDER*, has been statistically tested against C4.5 and C4.5 Rules, and performed well. The knowledge models obtained are simpler, with very few decision rules, and therefore easier to understand, which is an advantage in many domains. The experiments with high-dimensional datasets showed the same good behavior, maintaining the quality of the knowledge model with respect to prediction accuracy. Index Terms-Decision rules, evolutionary encoding, supervised learning. I. INTRODUCTION D ECISION RULES are especially relevant in problems related to supervised learning. Given a dataset with continuous and discrete features or attributes, and a class label, we try to find a rule set that describes the knowledge within data or classifies new unseen data. When the feature is discrete, the rules take the form of "if , then class," where the values are not necessarily all those that the feature can take. When the feature is continuous, typically the rules take the form of "if then class," where

Journal of Physics A: Mathematical and General, 1996

A formalism for describing the dynamics of Genetic Algorithms (GAs) using methods from statistical mechanics is applied to the problem of generalization in a perceptron with binary weights. The dynamics are solved for the case where a new batch of training patterns is presented to each population member each generation, which considerably simplifies the calculation. The theory is shown to agree closely to simulations of a real GA averaged over many runs, accurately predicting the mean best solution found. For weak selection and large problem size the difference equations describing the dynamics can be expressed analytically and we find that the effects of noise due to the finite size of each training batch can be removed by increasing the population size appropriately. If this population resizing is used, one can deduce the most computationally efficient size of training batch each generation. For independent patterns this choice also gives the minimum total number of training patterns used. Although using independent patterns is a very inefficient use of training patterns in general, this work may also prove useful for determining the optimum batch size in the case where patterns are recycled.

2012 IEEE Congress on Evolutionary Computation, 2012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. Highlights • Novelty Search is applied for the first time to supervised classification with GP. • Two new variants of NS are proposed, overcoming some of its main shortcomings. • NS achieves competitive results compared to objective-based search. • Results show bloat control properties in binary tasks for NS.

Proceedings of the 4th International Workshop on Artificial Neural Networks and Intelligent Information Processing, 2008

This paper aims at learning and evolution in artificial neural networks. Here is presented a system evolving populations of neural nets that are fully connected multilayer feedforward networks with fixed architecture solving given tasks. The system is compared with gradient descent weight training (like backpropagation) and with hybrid neural network adaptation. All neural networks have the same architecture and solve the same problems to be able to be compared mutually. In order to test the efficiency of described algorithms, we applied them to the Fisher's Iris data set [1] that is the bench test database from the area of machine learning.

2005

Proceedings of the 9th annual conference companion on Genetic and evolutionary computation, 2007

In this paper, we introduce an approach for the identification of building blocks in symbolic expressions and apply it to analyze the emergence of building blocks in XCS with symbolic representation. The objective is to extract from a sequence of evolving populations a set of recurrent patterns which identifies pieces of the problem solution, so to track the emergence of the optimal solution. This permits the introduction of better measures of performance which might be useful in diagnosing problems and adapting algorithms.

lsc.fie.umich.mx

Evolutionary Computation has been used to automatically design Artificial Neural Networks (ANN) capable of solving hard classification and forecasting problems. The evolutionary process contains a lot of information that may help in the design of future ANN for similar problems.

… , 2000. Proceedings of the …, 2000

Artificial neural networks (ANNs) are used extensively involving continuous data. However, their application in many domains is hampered because it is not clear how they partition continuous data for classification. The extraction of rules, therefore, from ANNs trained on continuous data is of great importance. The system described in this paper uses a genetic algorithm to generate input patterns which are presented to the network, and the output from the ANN is then used to calculate the fitness function for the algorithm. These ...