#Robins, A. and Frean, M. Local learning algorithms for sequential learning tasks in neural networks, Journal of Advanced Computational Intelligence, vol.2, 6 (1998)  Home/Search   Document Not in Database   Summary   Related Articles   Check

....node will be either updated, or created for each data example. This makes the learning procedure very fast (especially in the case when December,2001 15 specialised parallel hardware platforms are used) Another advantage is that learning a new data example does not cause forgetting of old ones [18,65]. A third advantage is that new input and new output variables can be added during the learning process, thus making the EFuNN system more flexible to accommodate new information, once such becomes available, without disregarding the already learned information. The use of MFs and membership ....

....may change over time in a continuous way. In this case the local on line error will depend on the closeness of the new input data to already stored prototypes in the existing rule nodes. The well established NN and AI techniques have difficulties when applied for on line, knowledge based learning [18,65,38]. For example, the multi layer perceptrons (MLP) and the backpropagation learning algorithm have the following problems: catastrophic forgetting [65,38] local minima problem [3,10] difficulties to extract rules [38] not able to adapt to new data without re training on old ones [65] too long ....

[Article contains additional citation context not shown here]

Robins, A. and Frean, M. "Local learning algorithms for sequential learning tasks in neural networks, Journal of Advanced Computational Intelligence, vol.2, 6 (1998)

....mode) rule node will be either updated, or created for each data example. This makes the learning procedure very fast (especially in the case when 14 specialised parallel hardware platforms are used) Another advantage is that learning a new data example does not cause forgetting of old ones [18,65]. A third advantage is that new input and new output variables can be added during the learning process, thus making the EFuNN system more flexible to accommodate new information, once such becomes available, without disregarding the already learned information. The use of MFs and membership ....

....may change over time in a continuous way. In this case the local on line error will depend on the closeness of the new input data to already stored prototypes in the existing rule nodes. The well established NN and AI techniques have difficulties when applied for on line, knowledge based learning [18,65,38]. For example, the multi layer perceptrons (MLP) and the backpropagation learning algorithm have the following problems: catastrophic forgetting [65,38] local minima problem [3,10] difficulties to extract rules [38] not able to adapt to new data without re training on old ones [65] too long ....

[Article contains additional citation context not shown here]

Robins, A. and Frean, M. "Local learning algorithms for sequential learning tasks in neural networks, Journal of Advanced Computational Intelligence, vol. 2, 6 (1998)

....to make decisions about its own improvement; to manifest introspection. 7. IS should adequately represent space and time in their different scales; should have parameters to represent such concepts as spatial distance, short term and long term memory, age, forgetting, etc. Several investigations [46, 10, 48]showed that the most popular neural network models and algorithms are not suitable for adaptive, on line learning. This includes multilayer perceptrons trained with the backpropagation algorithm, radial basis function networks, self organising maps SOMs [34, 35] and those NN models which were not ....

A. Robins and M. Freans. Local learning algorithms for sequential learning tasks in neural networks. Journal of Advanced Computational Intelligence, 2(6), 1998.

No context found.

#Robins, A. and Frean, M. Local learning algorithms for sequential learning tasks in neural networks, Journal of Advanced Computational Intelligence, vol.2, 6 (1998)

No context found.

Robins, A. and Frean, M. "Local learning algorithms for sequential learning tasks in neural networks, Journal of Advanced Computational Intelligence, vol.2, 6 (1998)

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC