these methods to high dimensional continuous state-spaces, which requires the use of function approximation to generalise the information learnt by the system. In this report, the use of back-propagation neural networks (Rumelhart, Hinton and Williams 1986) is considered in this context. We consider a number

Abstract – This presentation gives a brief introduction to Evolving Connectionist Systems (ECOS) and their applications in Bioinformatics, Brain study and Intelligent Machines. These systems evolve their structure and functionality through learning from data in both on-line and off-line incremental

1 (Bullinaria, 1999). Simulating brain lesions in connectionist systems was discussed by Small (1991). Bullinaria and Chater (1995) found that very similar patterns of deficits arose by randomly removing hidden units, randomly removing connections, globally scaling the weights, or adding random

this document were found in the intersection of the work of several insightful researchers. The generation and analysis of the adder networks which were described here was, for the most part, performed by FuSheng Tsung [3]. The Movie Description Network was the result of work by Brian Bartell [2]. Thanks are also due to Paul Churchland for his insights concerning the relationship between instruction sequences and regions of network activation space. REFERENCES

The main goal of this paper is to present a new learning algorithm which has been applied to feedforward neural networks. It was used not only for learning of a given network, but also to optimize the number of hidden layer neurons. Besides, this learning algorithm is inspired on the traditional backpropagation algorithm. Nevertheless it owns some variations due to kind of network used. This algorithm was applied to a particular network which has AND/OR fuzzy neurons.

inside the cognitive system which can be operated upon by structure sensitive rules similar to those employed in systems of natural deduction. I have argued elsewhere that the systematicity of human thought might better be explained as resulting from the fact that we have learned natural languages which

The paper presents a method for optimising parameter values of evolving connectionist systems (ECoS) for life-long learning. The method is based on evolutionary computation principles and on genetic algorithms in particular. The method is illustrated on a spoken phoneme data classi cation task. I.

The paper describes what evolving processes are and presents a computational model called evolving connectionist systems. The model is based on principles from both brain organization and genetics. The applicability of the model for dynamic modeling and knowledge discovery in the areas of brain

Key words : Neural networks, error estimation, novelty detection, validation. Connectionist systems, or neural networks, have made a significant impact in the world of statistical pattern classification, detection and prediction over the last few years. It is clear that the basic mechanisms

Abstract A method is presented for extending the Evolving Connectionist System (ECoS) algorithm that allows it to explic-itly represent and learn nominal-scale data without the need for an orthogonal or binary encoding scheme. Rigorous evaluation of the algorithm over benchmark data sets shows