Evolutionary Selection of Features for Neural Sleep/Wake Discrimination

Evolutionary Selection of Features for Neural Sleep/Wake Discrimination. Duerr, P., Karlen, W., Guignard, J., & Mattiussi, C. Journal of Artificial Evolution and Applications, 2009:1-10, 2009.

Website abstract bibtex

In biomedical signal analysis, artificial neural networks are often used for pattern classification because of their capability for nonlinear class separation and the possibility to efficiently implement them on amicrocontroller. Typically, the network topology is designed by hand, and a gradient-based search algorithm is used to find a set of suitable parameters for the given classification task. In many cases, however, the choice of the network architecture is a critical and difficult task. For example, hand-designed networks often requiremore computational resources than necessary because they rely on input features that provide no information or are redundant. In the case of mobile applications,where computational resources and energy are limited, this is especially detrimental. Neuroevolutionarymethods which allow for the automatic synthesis of network topology and parameters offer a solution to these problems. In this paper, we use analog genetic encoding (AGE) for the evolutionary synthesis of a neural classifier for a mobile sleep/wake discrimination system. The comparison with a hand-designed classifier trained with back propagation shows that the evolved neural classifiers display similar performance to the hand-designed networks, but using a greatly reduced set of inputs, thus reducing computation time and improving the energy efficiency of themobile system.

@article{
 title = {Evolutionary Selection of Features for Neural Sleep/Wake Discrimination},
 type = {article},
 year = {2009},
 identifiers = {[object Object]},
 pages = {1-10},
 volume = {2009},
 websites = {http://www.hindawi.com/journals/jaea/2009/179680.html},
 id = {a9e3b488-57ff-306f-8c76-036850715a17},
 created = {2009-05-29T19:40:48.000Z},
 file_attached = {true},
 profile_id = {6d353feb-efe4-367e-84a2-0815eb9ca878},
 last_modified = {2020-06-11T05:27:57.023Z},
 read = {true},
 starred = {false},
 authored = {true},
 confirmed = {true},
 hidden = {false},
 citation_key = {Duerr2009a},
 notes = {IF-2018: n/a},
 folder_uuids = {9aa961a4-04a6-4a65-b081-d534c7a81a35,4afa922c-d8d6-102e-ac9a-0024e85ead87},
 private_publication = {false},
 abstract = {In biomedical signal analysis, artificial neural networks are often used for pattern classification because of their capability for nonlinear class separation and the possibility to efficiently implement them on amicrocontroller. Typically, the network topology is designed by hand, and a gradient-based search algorithm is used to find a set of suitable parameters for the given classification task. In many cases, however, the choice of the network architecture is a critical and difficult task. For example, hand-designed networks often requiremore computational resources than necessary because they rely on input features that provide no information or are redundant. In the case of mobile applications,where computational resources and energy are limited, this is especially detrimental. Neuroevolutionarymethods which allow for the automatic synthesis of network topology and parameters offer a solution to these problems. In this paper, we use analog genetic encoding (AGE) for the evolutionary synthesis of a neural classifier for a mobile sleep/wake discrimination system. The comparison with a hand-designed classifier trained with back propagation shows that the evolved neural classifiers display similar performance to the hand-designed networks, but using a greatly reduced set of inputs, thus reducing computation time and improving the energy efficiency of themobile system.},
 bibtype = {article},
 author = {Duerr, P and Karlen, Walter and Guignard, J and Mattiussi, C},
 journal = {Journal of Artificial Evolution and Applications}
}

Downloads: 0