D. White, "GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design," Lecture Notes in Computer Science, vol. 686, pp. 322--327, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....i.e. one objective is to map abnormal observations to a value close to 0278 0062 99 10.00 1999 IEEE one and the other objective is to map normal observations to a value close to zero. Genetic algorithms (GA s) 9] have been applied to many diagnostic and classification problems [8] 10] [15]. A conventional GA, however, is a scalar optimization technique. It thus possesses the undesirable features of an aggregating based approach. One method of avoiding this is to adopt a multiobjective approach [16] 17] to the optimization problem. In a multiobjective optimization approach, the ....

D. White and P. Ligomenides, "GANNet: A genetic algorithm for optimizing topology and weights in neural network design," Lecture Notes Comput. Sci., no. 686, pp. 322--327, 1993.

.... to use of those evolutionary computation (EC) methods which have previously been applied to uni objective NN design, genetic algorithms (GAs) evolution strategies (ES) and particle swarm optimisation (PSO) GAs have previously be used for feature selection [8, 53] and topography selection [2, 5, 29, 35, 36, 38, 52] and ESs have been used for weight optimisation [21, 42, 45, 55] and adaptive topography selection [15, 37, 57] The recent EC technique of PSO [27] has also proved popular as a uni objective NN optimiser [10, 12, 13, 26, 48] 2 Multi objective evolutionary neural network flamework The use of ....

D. White. GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design. Lecture Notes in Computer Science, 686:322-327, 1993.

....of the connection from node to node . We can use to indicate a connection and to indicate no connection. In fact, can represent real valued connection weights from node to node so that the architecture and connection weights can be evolved simultaneously [37] 42] 45] 165] 166] 169] [171]. Each matrix has a direct one to one mapping to the corresponding ANN architecture. The binary string representing an architecture is the concatenation of rows (or columns) of the matrix. Constraints on architectures being explored can easily be incorporated into such a representation scheme by ....

....with seven nodes were considered. The transfer function was specified in the structural genes in their genotypic representation. It was much more complex than the usual sigmoid function because they tried to model a biological neuron in the tailflip circuitry of crayfish. White and Ligomenides [171] adopted a simpler approach to the evolution of both topological structures and node transfer functions. For each individual (i.e. ANN) in the initial population, 80 nodes in the ANN used the sigmoid transfer function and 20 nodes used the Gaussian transfer function. The evolution was used to ....

D. White and P. Ligomenides, "GANNet: A genetic algorithm for optimizing topology and weights in neural network design," in Proc. Int. Workshop Artificial Neural Networks (IWANN'93), Lecture Notes in Computer Science, vol. 686. Berlin, Germany: Springer-Verlag, 1993, pp. 322--327.

....that they concentrate on optimizing only a part of the neural net, disregarding the rest: the hidden layer size and the learning constants. Search over topology space: some authors use a GA whose population is a set of complete ANNs, initialized with di erent hidden layer sizes (White et al. in [27], Yao and Liu [5, 6] and Falco et al. 28] a method to prune oversized networks is also used (Bebis et al. 29] Miller et al. 7] propose a method in which the network connection structure is mapped onto a binary adjacency matrix called the Miller Matrix that describes the ANN architecture, ....

David White and Panos Ligomenides. GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design. IWANN93. Lectures Notes in Computer Science, vol. 686, 322-327, 1993.

....leads to poorer evolution than does an intermediate amount of learning. Main problem with this approach is that it concentrates on optimizing only a part of the neural net, disregarding the rest: hidden layer size and learning constants. Search over topology space, as did White et al. in [30], where a GA, using a population initialized with di erent hidden layer sizes, is presented. In this 3 approach, each individual is a complete neural network, and the allele is a hidden or output node with its associated input links. The main problem of this method is that it only search using ....

....capacities of the BP algorithm, such as generalization [43] Our opinion, along 8 with Prechelt [44] is that to test an algorithm, at least two real world problems should be used. This is why comparisons with other GA NN algorithms are not included in this paper: in other published papers [30,8,34,37], they were applied to toy or non public available problems. The tests were applied as follows: each data set was divided into three disjoint parts, for training, validating and testing. Thus, in order to obtain the tness of an individual, the MLP is trained with the training set and its tness ....

David White and Panos Ligomenides. GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design. Lectures Notes in Computer Science, Springer Verlag, vol. 686, 322-327, 1993.

....evolve a population of ANN by encoding the parameters of the hidden layer into binary strings. De Falco et al. propose a method [19] based on an evolutionary approach to provide the optimal set of synaptic weights of the network. Several authors search over topology space, as did White et al. in [20], Miller et al. in [6] or Bebis et al. 21] that proposes the couple between GAs and weight elimination. The methods proposed by Yao and Liu [5] and Castillo et al. 4] combine the search for the optimal set of weights and the search for the optimal topology, using a GA and BP. De Falco et al. ....

David White and Panos Ligomenides. GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design. IWANN93. Lectures Notes in Computer Science, vol. 686, 322-327, 1993.

....a population of ANN by encoding the parameters of the hidden layer into binary strings. De Falco et al. propose a method [15] based on an evolutionary approach to provide the optimal set of synaptic weights of the network. ffl Several authors search over topology space, as did White et al. in [16], Miller et al. in [17] or Bebis et al. 18] that proposes the couple between GAs and weight elimination. The methods proposed by Yao and Liu [19] and Castillo et al. 20] combine the search for the optimal set of weights and the search for the optimal topology, using a GA and BP. De Falco et al. ....

David White and Panos Ligomenides. GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design. IWANN93. Lectures Notes in Computer Science, vol. 686, 322-327, 1993.

....parents from the population based on their fitness. 5. Apply search operators to the parents and generate offspring which form the next generation. Figure 6: A typical cycle of the evolution of architectures. Considerable research on evolving ANN architectures has been carried out in recent years [33, 42, 45, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 149, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 138, 213, 214, 215, 216, 118, 130, 127, 217, 218, 219, 220, 221, 222, 223, 128, 224, 225]. Most of the research has concentrated on the evolution of ANN topological structures. Relatively little has been done on the evolution of node transfer functions, let al..one the simultaneous evolution of both topological structures and node transfer functions. In this paper, we will analyze the ....

....or absence of the connection from node i to node j. We can use c ij = 1 to indicate a connection and c ij = 0 to indicate no connection. In fact, c ij can represent real valued connection weights from node i to node j so that the architecture and connection weights can be evolved simultaneously [165, 171, 169, 170, 166, 42, 45, 37]. Each matrix C has a direct one to one mapping to the corresponding ANN architecture. The binary string representing an architecture is the concatenation of rows (or columns) of the matrix. Constraints on architectures being explored can easily be incorporated into such a representation scheme by ....

[Article contains additional citation context not shown here]

D. White and P. Ligomenides, "GANNet: a genetic algorithm for optimizing topology and weights in neural network design," in Proc. of Int'l Workshop on Artificial Neural Networks (IWANN'93), pp. 322--327, Springer-Verlag, 1993. Lecture Notes in Computer Science, Vol. 686.

....the next generation. Figure 6: A typical cycle of the evolution of architectures. Reprinted with permission from Ref. 1. X. Yao: Evolutionary Artificial Neural Networks 19 Because of advantages of the evolutionary design of architectures, a lot of research has been carried out in recent years [37, 46, 49, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100]. However, almost all the research only deals with the topological structure of EANNs and little has been done on the evolution of node transfer functions let al..one the evolution of both topological structures and node transfer functions. We will concentrate on the evolution of topological ....

....or absence of the connection from node i to node j. We can use c ij = 1 to indicate a connection and c ij = 0 no connection. In fact, c ij can even be connection weights from node i to node j so that both the topological structure and connection weights of an EANN are evolved at the same time [93, 99, 97, 98, 94, 46, 49, 41]. Each such matrix has a direct one to one mapping to the corresponding architecture. The binary string representing an architecture is just the concatenation of rows (or columns) of the matrix. Constraints on architectures being explored can easily be incorporated into such representation scheme ....

[Article contains additional citation context not shown here]

D. White and P. Ligomenides. GANNet: a genetic algorithm for optimizing topology and weights in neural network design. In Proc. of Int'l Workshop on Artificial Neural Networks (IWANN'93), pages 322--327. Springer-Verlag, 1993. Lecture Notes in Computer Science, Vol. 686.

....the children generated above, and obtain the next generation. Figure 3: A typical cycle of the evolution of architectures. X. Yao: Evolutionary Artificial Neural Networks 13 Because of advantages of the evolutionary design of architectures, a lot of research has been carried out in recent years [36, 45, 48, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96]. However, almost all the research only deals with the topological structure of EANNs, little has been done on the evolution of node transfer functions, let al..one the evolution of both topological structures and node transfer functions. We will concentrate on the evolution of topological ....

....or absence of the connection from node i to node j. We can use c ij = 1 to indicate a connection and c ij = 0 no connection. In fact, c ij can even be connection weights from node i to node j so that both the topological structure and connection weights of an EANN are evolved at the same time [89, 95, 93, 94, 90, 45, 48, 40]. Each such matrix has a direct one to one mapping to the corresponding architecture. The binary string representing an architecture is just the concatenation of rows (or columns) of the matrix. Constraints on architectures being explored can easily be incorporated into such representation scheme ....

[Article contains additional citation context not shown here]

D. White and P. Ligomenides. GANNet: a genetic algorithm for optimizing topology and weights in neural network design. In Proc. of Int'l Workshop on Artificial Neural Networks (IWANN'93), pages 322--327. Springer-Verlag, 1993. Lecture Notes in Computer Science, Vol. 686.

No context found.

D. White, "GANNet: A Genetic Algorithm for Optimizing Topology and Weights in Neural Network Design," Lecture Notes in Computer Science, vol. 686, pp. 322--327, 1993.

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