Hugo de Gads, "Genetic Programming", Ch.8 in book Neural and IntelligentSystems Integration, ed. Branko Soucek, Wiley, NY, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....unfortunately encoding strategies with high modularity like indirect encoding, have not been a breakthrough. Another idea is to partially develop NNs for subtasks and compose them to a more complex structure. This was investigated for different parts of the nervous system of an artificial being [2, 3]. At this stage, however, the efficient application of GANN systems seems to be restricted to (a) weight training problems where no error information is available, such as for the pole balancing problem, or the recurrent networks, and (b) architecture optimization for classes of problems that ....

Hugo de Garis. Genetic programming. In International Joint Conference on Neural Networks, pages III 511--516. IEEE, 1990.

....Friedberg, R. M. 491, 492] Fuchs, Dirk, 428] Fuchs, Marc, 428] Fuchs, Mathias, 462] Fuchs, Matthias, 428] Fujiki, Cory, 488] Fuller, Stephanie R. 414] Furuya, Tatsumi, 539] Gamble, Rose, 292] Garces Perez, Jaime, 284] Garcia, O. N. 190] Garcia, Oscar N. 429] Garis, Hugo de, [150, 165, 172, 202, 536, 537, 538, 539, 540, 541] Gathercole, Chris, 96, 285, 460] Geyer Schulz, Andreas, 406] Ghanea Hercock, R. 97] Gibbs, Jonathan, 302] Gibbs, W. Wayt, 342] Glevarec, P. 187] Gofman, Yossi, 314] Goldfish, Andrew, 286] Goodman, Erik D. 326] Gordon, Benjamin M. 84] Gray, G. J. 353] Gray, H. F. 287, 409] ....

.... programming, 104] fitting Mackey Glass, 119] formal languages context free, 248] fractals IFS, 187] ftiness limited error, 460] function approximation, 180] GA P, 200] game theory, 279] games, 390] Nim, 422] poker, 392] Tetris, 234] tile puzzle, 188] genetic programming, [489, 485, 490, 506, 483, 487, 488, 37, 507, 38, 39, 41, 42, 43, 44, 45, 46, 47, 48, 49, 474, 475, 501, 508, 40, 50, 51, 52, 53, 54, 55, 56, 57, 58, 525, 530, 536, 537, 77, 493, 494, 499, 500, 509, 511, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 519, 520, 522, 524, 526, 527, 538, 539, 476, 477, 478, 479, 480, 481, 482, 484, 486, 495, 496, 497, 498, 502, 503, 504, 505, 510, 512, 514, 71, 72, 73, 74, 75, 515, 516, 517, 518, 521, 523, 528, 529, 531, 532, 533, 534, 540, 541, 79, 82, 84, 86, 87, 88, 89, 92, 95, 96, 97, 99, 102, 107, 108, 109, 112, 114, 115, 117, 122, 123, 124, 125, 127, 129, 130, 133, 134, 135, 137, 138, 139, 140, 142, 143, 147, 149, 153, 155, 156, 157, 158, 160, 164, 167, 168, 170, 171, 10, 173, 11, 12, 13, 16, 17, 18, 19, 20, 175, 176, 177, 178, 180, 181, 182, 183, 184, 185, 187, 188, 189, 190, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 205, 21, 206, 207, 210, 211, 213, 214, 215, 216, 217, 218, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 233, 234, 236, 237, 22, 238, 239, 240, 241, 242, 244, 245, 246, 248, 249, 250, 252, 253, 254, 255, 257, 258, 259, 260, 307, 261, 262, 26, 264, 265, 266, 267, 270, 272, 275, 277, 278, 279, 280, 281, 283, 284, 288, 289, 290, 292, 293, 294, 296, 298, 301, 302, 303, 304, 305, 306, 308, 309, 310, 27, 313, 314, 315, 316, 317, 318, 320, 321, 322, 323, 328, 330, 331, 332, 333, 334, 335, 336, 337, 338, 339, 341, 343, 344, 345, 346, 347, 348, 28, 349, 350, 351, 352, 355, 357, 358, 359, 360, 29, 361, 362, 363, 364, 366, 367, 368, 371, 372, 373, 374, 375, 32, 376, 377, 378, 379, 380, 381, 382, 383, 384, 385, 386, 387, 388, 390, 391, 392, 393, 394, 395, 396, 397, 400, 402, 403, 404, 405, 407, 408, 409, 410, 411, 412, 413, 414, 415, 417, 421, 422, 423, 33, 425, 426, 427, 429, 430, 432, 433, 435, 437, 438, 439, 440, 442, 443, 444, 445, 446, 447, 448, 449, 450, 451, 453, 454, 455, 456, 459, 461, 34, 462, 463, 464, 465, 466, 467, 468, 35, 469, 470, 471, 36, 472] genetic programming acyclic graphs, 151] agents, 325] AI, 329] analysis, 159, 327] Boolean functions, 398, 460] breeding, 80, 169, 172] C, 434] C , 103, 128, 186, 354, 356] classification, 287] code reuse, 418] combinatorial logic, 428] commercial applications, 243] compact ....

[Article contains additional citation context not shown here]

Hugo de Garis. Genetic programming. In Branko Soucek, editor, Neural and Intelligent Systems Integration, chapter 8. Joh Wiley & Sons, New York, 1991. y ga:deGaris91d.

....levels will be so high (especially when nanotechnology (i.e. molecular scale technology) becomes a reality) that no human being will be able to predict or even analyze how these systems function. The concept of evolutionary building of complex systems is called Genetic Programming (GP) [1,2,3,4]. This paper shows how artificial neural networks, based on cellular automata can be grown, using GP techniques. The ideas and results of this project will serve as the conceptual basis for the construction of what are called Darwin Machines [2] A Darwin Machine is a special hardware device ....

....systems is called Genetic Programming (GP) 1,2,3,4] This paper shows how artificial neural networks, based on cellular automata can be grown, using GP techniques. The ideas and results of this project will serve as the conceptual basis for the construction of what are called Darwin Machines [2]. A Darwin Machine is a special hardware device used to perform GP in parallel. For example, Cellular Automata Machines could function in parallel to evolve the neurite networks described below. Each CAM would have a conventional programmable processor to measure the fitness of the evolved neurite ....

[Article contains additional citation context not shown here]

Hugo de Garis, "Genetic Programming", Ch.8 in book "Neural and Intelligent Systems Integration", ed. Branko Soucek, WILEY, 1991.

....Patrik, 66, 66] Dickinson, Andrew, 93] Dickinson, John, 144] Dosi, G. 43] Dunay, Bertrand Daniel, 67, 110] Durnota, Bohdan, 186] Forsyth, Richard S. 16, 187] Franguiadakis, Terry, 65] Fraser, A. P. 188, 189, 181] Fujiki, Cory, 144] Furuya, Tatsumi, 164] Garis, Hugo de, [89, 83, 190, 112, 130, 191, 192, 164, 169, 34] Gathercole, C. 85] Ghanea Hercock, R. 189] Gordon, Benjamin M. 91] Gruau, Fr ed eric C. 193, 21, 31, 125, 149, 194, 32, 15] Hampo, Richard J. 182, 195] Handley, Simon G. 68, 81, 196, 197, 150, 198, 199, 200] Harala, Sauli, 201] Hasegawa, Yoshishige, 176] Haynes, Thoms, ....

.... [101] trade strategies, 94] trading, 173] editorial, 210] engineering automobile, 182] ethology territory defining behavior in birds, 97] evolution, 123, 11, 39] filters stack, 233] fitness landscapes genetic programming, 69] fitting Mackey Glass, 233] genetic programming, [16, 140, 187, 24, 142, 143, 144, 217, 138, 206, 42, 166, 50, 218, 219, 220, 51, 207, 54, 55, 26, 121, 197, 30, 161, 221, 222, 146, 127, 123, 134, 223, 224, 46, 14, 129, 130, 191, 201, 31, 125, 203, 195, 230, 22, 225, 226, 10, 227, 228, 136, 48, 49, 11, 39, 47, 53, 52, 163, 40, 175, 25, 128, 238, 27, 192, 164, 177, 44, 29, 179, 148, 36, 17, 168, 132, 149, 194, 32, 15, 150, 198, 199, 200, 205, 151, 158, 159, 152, 208, 229, 43, 28, 153, 35, 18, 37, 20, 8, 154, 155, 242, 245, 169, 34, 7, 234, 173, 178, 91, 92, 184, 93, 188, 85, 189, 182, 204, 94, 95, 231, 96, 202, 87, 186, 76, 181, 97, 237, 239, 240, 99, 78, 100, 101, 102, 57, 58, 88, 21, 38, 45, 170, 243, 172, 180, 244, 19, 23, 212, 60, 9, 215, 41, 216, 120, 108, 109, 110, 111, 112, 104, 113, 114, 115, 105, 116, 117, 118] genetic programming acyclic graphs, 81] analysis, 171] breeding, 235, 241, 190] C , 209, 13, 12] control, 82] crossover, 66] databases, 71] distributed, 98] donut problem, 241] double based, 183] dynamical systems, 72] editorial, 210] emergent phenomena, 174] filters, ....

[Article contains additional citation context not shown here]

Hugo de Garis. Genetic programming. In Branko Soucek, editor, Neural and Intelligent Systems Integration, chapter 8. Joh Wiley & Sons, New York, 1991. y ga:deGaris91d.

....EANNs which are less fit (have large errors) The main idea here is to use GAs as function optimisers to maximise fitness functions (minimise error functions) since GAs are good at dealing with large, complex, nondifferentiable and deceptive spaces. A lot of work has been done along this line [24, 25, 26, 27, 28, 29, 30, 31, 32, 33]. The evolutionary training approach is divided into two major steps; the first one is to decide the representation scheme of connection weights, e.g. whether it is in the form of binary strings or not, and the second one is the evolution itself driven by GAs. Different representation schemes and ....

H. de Garis. Genetic programming. In Proc. of Int'l Joint Conf. on Neural Networks, Vol. I, pages 194--197, Washington, DC, 1990. Lawrence Erlbaum Associates, Hillsdale, NJ.

....left, turn right, peck at food and mate [6] Each of these behaviors was controlled by the time varying outputs of a single evolved neural network module, and applied to the angles of the leg components of LIZZY. As far as he is aware, the author was the first person to evolve neural net dynamics [3], in the form of walking stick legs Walker ) Switching between behaviors involved taking the outputs from one neural net module and feeding them into the inputs of the next module. The next step was to evolve neural net detectors, e.g. for frequency, signal strength, signal strength ....

Hugo de Gads, "Genetic Programming", Ch.8 in book Neural and IntelligentSystems Integration, ed. Branko Soucek, Wiley, NY, 1991.

....: The LIZZY Project 9. Future Ideas for The LIZZY Project 10. Genetic Programming (GP) of Embryos 11. Future Embryo Ideas : Electronic Neuro Embryology 12. References 1. Introduction This article introduces the concept of Genetic Programming in a more general way than in earlier papers [2,3,4,5]. Genetic Programming (GP) is the application of the Genetic Algorithm [7,8] to the creation of systems which are too complex in their dynamics to be analyzed or pre specified in detail. Such systems can be built, but (probably) not understood. Two major applications of GP will be introduced in ....

....Neural Networks and Artificial Life. Many Lifers (Artificial Life researchers) hope that their bottom up behavioural approach and the top down disembodied symbolic approach of AI researchers will one day meet somewhere in the middle. 10. Genetic Programming (GP) of Embryos Earlier papers [2,3,4,5] defined Genetic Programming (GP) as a new programming methodology which used the Genetic Algorithm to evolve neural network modules. However, the ideas of this section and the next will show that the GP approach is quite general. For example, it can be applied to the evolution of artificial ....

H. de Garis, "Genetic Programming", Chapter 15, in book, "Neural and Intelligent Systems Integration", ed. Prof. Branko Soucek, Wiley Interscience, 1991, (to appear).

....or even analyzable. Therefore, in order to build such an artificial brain, an evolutionary engineering approach will be used which the author calls Genetic Programming (GP) i.e. using evolutionary algorithms such as the Genetic Algorithm [6] as tools to build grow evolve complex systems [1,2,3,4,8]. Having decided to use GP to build an artificial brain, the next step was to decide to perform the evolution at electronic speeds, in order to accelerate the whole process. This lead to the concept of the Darwin Machine [8] which is special hardware which performs GP in parallel. However, the ....

....at the moment is whether, after all the 10,000 rule effort is completed, these CA based neural networks will prove to be evolvable, i.e. will their fitness levels continue to increase to the point where their functionality reaches desired levels. The author has already experimented with GenNets [1,2,3,4], i.e. fully connected software simulated neural nets, whose N N weights (for an N neuron module) were concatenated onto a GA chromosome and evolved. These GenNets proved to be highly evolvable, as well as being robust. In fact, one could often cut (i.e. make zero) 70 of the weights, and still ....

[Article contains additional citation context not shown here]

Hugo de Garis, "Genetic Programming", Ch.8 in book "Neural and Intelligent Systems Integration", ed. Branko Soucek, WILEY, 1991.

....left, turn right, peck at food and mate [6] Each of these behaviors was controlled by the time varying outputs of a single evolved neural network module, and applied to the angles of the leg components of LIZZY. As far as he is aware, the author was the first person to evolve neural net dynamics [3], in the form of walking stick legs Walker ) Switching between behaviors involved taking the outputs from one neural net module and feeding them into the inputs of the next module. The next step was to evolve neural net detectors, e.g. for frequency, signal strength, signal strength ....

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and IntelligentSystems Integration, ed. Branko Soucek, Wiley, NY, 1991.

....peck at food and mate [de Garis 1994] Each of these behaviors was controlled by the time varying outputs of a single evolved neural network module, and applied to the angles of the leg components of LIZZY. As far as he is aware, the author was the first person to evolve neural net dynamics [de Garis 1991], in the form of walking stick legs Walker ) Switching between behaviors involved taking the outputs from one neural net module and feeding them into the inputs of the next module. The next step was to evolve neural net detectors, e.g. for frequency, signal strength, signal strength ....

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and ntelligent Systems Integration, ed. Branko Soucek, Wiley, NY, 1991.

....neurons in various directions. The above experiments are only the beginning. The author has already evolved (not using CAs) the weights of recurrent neural networks as controllers of an artificial nervous system for a simulated quadruped artificial creature. Neural modules called GenNets [de Garis 1990, 1991, 1994] were evolved to make the creature walk straight, turn left or right, peck at food, and mate. GenNets were also evolved to detect signal frequencies, to generate signal frequencies, to detect signal strengths, and signal strength differences. By using the output of the detector GenNets, it ....

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and Intelligent Systems Integration, ed. Branko Soucek, Wiley, NY, 1991.

No context found.

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and Intelligent Systems Integration, ed. Branko Soucek, Wiley, NY, 1991.

....when nanotechnology (i.e. molecular scale technology) becomes a reality) that no human being will be able to predict or even analyze how these systems function. The author has given the concept of evolutionary building of complex systems a label. He calls it Genetic Programming (GP) [1,2,3,4]. This paper shows how an artificial neural network, based on cellular automata can be grown, using GP techniques. The ideas and results of this paper will serve as the conceptual basis for the construction of what the author calls Darwin Machines [2] A Darwin Machine is a special hardware ....

....He calls it Genetic Programming (GP) 1,2,3,4] This paper shows how an artificial neural network, based on cellular automata can be grown, using GP techniques. The ideas and results of this paper will serve as the conceptual basis for the construction of what the author calls Darwin Machines [2]. A Darwin Machine is a special hardware device used to perform GP in parallel. For example, Cellular Automata Machines could function in parallel to evolve the neurite networks described below. Each CAM would have a conventional programmable processor to measure the fitness of the evolved neurite ....

[Article contains additional citation context not shown here]

) Hugo de Garis, "Genetic Programming", Ch.8 in book "Neural and Intelligent Systems Integration", ed. Branko Soucek, WILEY, 1991.

....The need for greater speed is obvious. The above experiments are only the beginning. The author has already evolved (not using CAs) the weights of recurrent neural networks as controllers of an artificial nervous system for a simulated quadruped artificial creature. Neural modules called GenNets [de Garis 1990, 1991, 1994] were evolved to make the creature walk straight, turn left or right, peck at food, and mate. GenNets were also evolved to detect signal frequencies, to generate signal frequencies, to detect signal strengths, and signal strength differences. By using the output of the detector GenNets, it ....

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and Intelligent Systems Integration, ed. Branko Soucek, Wiley, NY, 1991.

....or even analyzable. Therefore, in order to build such an artificial brain, an evolutionary engineering approach will be used which the author calls Genetic Programming, GP) i.e. using evolutionary algorithms such as the genetic algorithms [6] as tools to build grow evolve complex systems [1,2,3,4,8]. Having decided to use GP to build an artificial brain, the next step was to decide to perform the evolution at electronic speeds, in order to accelerate the whole process. This lead to the concept of the Darwin Machine [8] which is special hardware which performs GP in parallel. However, the ....

....at the moment is whether, after all the 10,000 rule effort is completed, these CA based neural networks will prove to be evolvable, i.e. will their fitness levels continue to increase to the point where their functionality reaches desired levels. The author has already experimented with GenNets [1,2,3,4], i.e. fully connected software simulated neural nets, whose N N weights (for an N neuron module) were concatenated onto a GA chromosome and evolved. These GenNets proved to be highly evolvable, as well as being robust. In fact, it was often possible to cut (i.e. make zero) 70 of the weights, ....

[Article contains additional citation context not shown here]

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and Intelligent Systems Integration, ed. Branko Soucek, Wiley, NY, 1991.

....need for greater speed is obvious. The above experiments are only the beginning. The author has already evolved (not using CAs) the weights of recurrent neural networks as controllers of an artificial nervous system for a simulated quadrupedal artificial creature. Neural modules called GenNets [de Garis 1991] were evolved to make the creature walk straight, turn left or right, peck at food, and mate. GenNets were also evolved to detect signal frequencies, to generate signal frequencies, to detect signal strengths, and signal strength differences. By using the output of the detector GenNets, it was ....

Hugo de Garis, "Genetic Programming", Ch.8 in book Neural and Intelligent Systems I ntegration, ed. Branko Soucek, Wiley, NY, 1991.

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