Automatic symbolic modelling of co-evolutionarily learned robot skills. Ledezma, A., Berlanga, A., & Aler, R. In Mira, J. & Prieto, A., editors, Connectionist Models of Neurons, Learning Processes and Artificial Intelligence, 6th International Work-Conference on Artificial and Natural Neural Networks, IWANN 2001, Proceedings, Part I, volume 2084, of Lecture Notes In Computer Science, pages 799–806, Granada, Spain, June, 2001. Springer. abstract bibtex Evolutionary based learning systems have proven to be very powerful techniques for solving a wide range of tasks, from prediction to optimization. However, in some cases the learned concepts are unreadable for humans. This prevents a deep semantic analysis of what has been really learned by those systems. We present in this paper an alternative to obtain symbolic models from subsymbolic learning. In the first stage, a subsymbolic learning system is applied to a given task. Then, a symbolic classifier is used for automatically generating the symbolic counterpart of the subsymbolic model. We have tested this approach to obtain a symbolic model of a neural network. The neural network defines a simple controller of an autonomous robot. A competitive coevolutive method has been applied in order to learn the right weights of the neural network. The results show that the obtained symbolic model is very accurate in the task of modelling the subsymbolic system, adding to this its readability characteristic.
@INPROCEEDINGS{modeling-iwann01,
author = {Agapito Ledezma and Antonio Berlanga and Ricardo Aler},
title = {Automatic symbolic modelling of co-evolutionarily learned robot skills},
booktitle = {Connectionist Models of Neurons, Learning Processes and Artificial
Intelligence, 6th International Work-Conference on Artificial and
Natural Neural Networks, IWANN 2001, Proceedings, Part I},
year = {2001},
editor = {José Mira and Alberto Prieto},
volume = {2084},
series = {Lecture Notes In Computer Science},
pages = {799--806},
address = {Granada, Spain},
month = {June},
publisher = {Springer},
abstract = {Evolutionary based learning systems have proven to be very powerful
techniques for solving a wide range of tasks, from prediction to
optimization. However, in some cases the learned concepts are unreadable
for humans. This prevents a deep semantic analysis of what has been
really learned by those systems. We present in this paper an alternative
to obtain symbolic models from subsymbolic learning. In the first
stage, a subsymbolic learning system is applied to a given task.
Then, a symbolic classifier is used for automatically generating
the symbolic counterpart of the subsymbolic model. We have tested
this approach to obtain a symbolic model of a neural network. The
neural network defines a simple controller of an autonomous robot.
A competitive coevolutive method has been applied in order to learn
the right weights of the neural network. The results show that the
obtained symbolic model is very accurate in the task of modelling
the subsymbolic system, adding to this its readability characteristic.},
bib2html_pubtype = {Refereed Conference},
bib2html_rescat = {Agent Modeling},
days = {13--15},
isbn = {3-540-42235-8},
key = {modelado},
owner = {ledezma},
timestamp = {2011.11.21}
}
Downloads: 0
{"_id":{"_str":"5342a9600e946d920a002fbc"},"__v":2,"authorIDs":["54576e722abc8e9f370003c4","8mdzoMZ27yguKiBHc"],"author_short":["Ledezma, A.","Berlanga, A.","Aler, R."],"bibbaseid":"ledezma-berlanga-aler-automaticsymbolicmodellingofcoevolutionarilylearnedrobotskills-2001","bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["Agapito"],"propositions":[],"lastnames":["Ledezma"],"suffixes":[]},{"firstnames":["Antonio"],"propositions":[],"lastnames":["Berlanga"],"suffixes":[]},{"firstnames":["Ricardo"],"propositions":[],"lastnames":["Aler"],"suffixes":[]}],"title":"Automatic symbolic modelling of co-evolutionarily learned robot skills","booktitle":"Connectionist Models of Neurons, Learning Processes and Artificial Intelligence, 6th International Work-Conference on Artificial and Natural Neural Networks, IWANN 2001, Proceedings, Part I","year":"2001","editor":[{"firstnames":["José"],"propositions":[],"lastnames":["Mira"],"suffixes":[]},{"firstnames":["Alberto"],"propositions":[],"lastnames":["Prieto"],"suffixes":[]}],"volume":"2084","series":"Lecture Notes In Computer Science","pages":"799–806","address":"Granada, Spain","month":"June","publisher":"Springer","abstract":"Evolutionary based learning systems have proven to be very powerful techniques for solving a wide range of tasks, from prediction to optimization. However, in some cases the learned concepts are unreadable for humans. This prevents a deep semantic analysis of what has been really learned by those systems. We present in this paper an alternative to obtain symbolic models from subsymbolic learning. In the first stage, a subsymbolic learning system is applied to a given task. Then, a symbolic classifier is used for automatically generating the symbolic counterpart of the subsymbolic model. We have tested this approach to obtain a symbolic model of a neural network. The neural network defines a simple controller of an autonomous robot. A competitive coevolutive method has been applied in order to learn the right weights of the neural network. The results show that the obtained symbolic model is very accurate in the task of modelling the subsymbolic system, adding to this its readability characteristic.","bib2html_pubtype":"Refereed Conference","bib2html_rescat":"Agent Modeling","days":"13–15","isbn":"3-540-42235-8","key":"modeling-iwann01","owner":"ledezma","timestamp":"2011.11.21","bibtex":"@INPROCEEDINGS{modeling-iwann01,\r\n author = {Agapito Ledezma and Antonio Berlanga and Ricardo Aler},\r\n title = {Automatic symbolic modelling of co-evolutionarily learned robot skills},\r\n booktitle = {Connectionist Models of Neurons, Learning Processes and Artificial\r\n\tIntelligence, 6th International Work-Conference on Artificial and\r\n\tNatural Neural Networks, IWANN 2001, Proceedings, Part I},\r\n year = {2001},\r\n editor = {José Mira and Alberto Prieto},\r\n volume = {2084},\r\n series = {Lecture Notes In Computer Science},\r\n pages = {799--806},\r\n address = {Granada, Spain},\r\n month = {June},\r\n publisher = {Springer},\r\n abstract = {Evolutionary based learning systems have proven to be very powerful\r\n\ttechniques for solving a wide range of tasks, from prediction to\r\n\toptimization. However, in some cases the learned concepts are unreadable\r\n\tfor humans. This prevents a deep semantic analysis of what has been\r\n\treally learned by those systems. We present in this paper an alternative\r\n\tto obtain symbolic models from subsymbolic learning. In the first\r\n\tstage, a subsymbolic learning system is applied to a given task.\r\n\tThen, a symbolic classifier is used for automatically generating\r\n\tthe symbolic counterpart of the subsymbolic model. We have tested\r\n\tthis approach to obtain a symbolic model of a neural network. The\r\n\tneural network defines a simple controller of an autonomous robot.\r\n\tA competitive coevolutive method has been applied in order to learn\r\n\tthe right weights of the neural network. The results show that the\r\n\tobtained symbolic model is very accurate in the task of modelling\r\n\tthe subsymbolic system, adding to this its readability characteristic.},\r\n bib2html_pubtype = {Refereed Conference},\r\n bib2html_rescat = {Agent Modeling},\r\n days = {13--15},\r\n isbn = {3-540-42235-8},\r\n key = {modelado},\r\n owner = {ledezma},\r\n timestamp = {2011.11.21}\r\n}\r\n\r\n","author_short":["Ledezma, A.","Berlanga, A.","Aler, R."],"editor_short":["Mira, J.","Prieto, A."],"id":"modeling-iwann01","bibbaseid":"ledezma-berlanga-aler-automaticsymbolicmodellingofcoevolutionarilylearnedrobotskills-2001","role":"author","urls":{},"downloads":0,"html":"","metadata":{"authorlinks":{"ledezma, a":"http://www.caos.inf.uc3m.es/ledezma/ca/publicaciones/"}}},"bibtype":"inproceedings","biburl":"http://www.caos.inf.uc3m.es/bibs/pubCAOS.bib","downloads":0,"keywords":[],"search_terms":["automatic","symbolic","modelling","evolutionarily","learned","robot","skills","ledezma","berlanga","aler"],"title":"Automatic symbolic modelling of co-evolutionarily learned robot skills","year":2001,"dataSources":["Z4T4fewybNpYCWveD"]}