%0 Conference Paper %1 zhang:evows05 %A Zhang, Mengjie %A Smart, Will %B Applications of Evolutionary Computing, EvoWorkshops2005: EvoBIO, EvoCOMNET, EvoHOT, EvoIASP, EvoMUSART, EvoSTOC %C Lausanne, Switzerland %D 2005 %E Rothlauf, Franz %E Branke, Juergen %E Cagnoni, Stefano %E Corne, David W. %E Drechsler, Rolf %E Jin, Yaochu %E Machado, Penousal %E Marchiori, Elena %E Romero, Juan %E Smith, George D. %E Squillero, Giovanni %I Springer Verlag %K algorithms, computation evolutionary genetic programming, %P 417--427 %R doi:10.1007/b106856 %T Learning Weights in Genetic Programs Using Gradient Descent for Object Recognition %V 3449 %X the use of gradient descent search in tree based genetic programming for object recognition problems. A weight parameter is introduced to each link between two nodes in a program tree. The weight is defined as a floating point number and determines the degree of contribution of the sub-program tree under the link with the weight. Changing a weight corresponds to changing the effect of the sub-program tree. The weight changes are learnt by gradient descent search at a particular generation. The programs are evolved and learned by both the genetic beam search and the gradient descent search. This approach is examined and compared with the basic genetic programming approach without gradient descent on three object classification problems of varying difficulty. The results suggest that the new approach works well on these problems. %@ 3-540-25396-3

@inproceedings{zhang:evows05, abstract = {the use of gradient descent search in tree based genetic programming for object recognition problems. A weight parameter is introduced to each link between two nodes in a program tree. The weight is defined as a floating point number and determines the degree of contribution of the sub-program tree under the link with the weight. Changing a weight corresponds to changing the effect of the sub-program tree. The weight changes are learnt by gradient descent search at a particular generation. The programs are evolved and learned by both the genetic beam search and the gradient descent search. This approach is examined and compared with the basic genetic programming approach without gradient descent on three object classification problems of varying difficulty. The results suggest that the new approach works well on these problems.}, added-at = {2008-06-19T17:46:40.000+0200}, address = {Lausanne, Switzerland}, author = {Zhang, Mengjie and Smart, Will}, biburl = {https://www.bibsonomy.org/bibtex/2528734f8be01b62e9fea1bad8d193cce/brazovayeye}, booktitle = {Applications of Evolutionary Computing, EvoWorkshops2005: {EvoBIO}, {EvoCOMNET}, {EvoHOT}, {EvoIASP}, {EvoMUSART}, {EvoSTOC}}, doi = {doi:10.1007/b106856}, editor = {Rothlauf, Franz and Branke, Juergen and Cagnoni, Stefano and Corne, David W. and Drechsler, Rolf and Jin, Yaochu and Machado, Penousal and Marchiori, Elena and Romero, Juan and Smith, George D. and Squillero, Giovanni}, interhash = {8235b4b70e55d6045a28a9ccafb0750e}, intrahash = {528734f8be01b62e9fea1bad8d193cce}, isbn = {3-540-25396-3}, issn = {0302-9743}, keywords = {algorithms, computation evolutionary genetic programming,}, month = {30 March-1 April}, notes = {EvoWorkshops2005}, pages = {417--427}, publisher = {Springer Verlag}, publisher_address = {Berlin}, series = {LNCS}, timestamp = {2008-06-19T17:55:35.000+0200}, title = {Learning Weights in Genetic Programs Using Gradient Descent for Object Recognition}, volume = 3449, year = 2005 }