This paper discusses the empirical evaluation of improving generalization performance of neural networks by systematic treatment of training and test failures. As a result of syst...
We present a design for an automated theorem prover that controls its search based on ideas from several areas of artificial intelligence (AI). The combination of case-based reaso...
of it suffer from a number of limitations, including: The effectiveness of an analogical reasoner depends upon its ability to select a relevant analogical source. In many problem d...
Abstract - The paper describes how a formal model of selfadaptation [Angeline, 1995] can be expressed in terms of Cultural Algorithms. A particular form of self-adaptation concerns...
Abstract—In this work we present a variational formulation for a multilayer perceptron neural network. With this formulation any learning task for the neural network is defined ...