Artificial Evolution Methods in the Biological and Biomedical SciencesView this Special Issue
Evolutionary Selection of Features for Neural Sleep/Wake Discrimination
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 a microcontroller. 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 require more 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. Neuroevolutionary methods 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 the mobile system.
A. Rechtschaffen, A. Kales, R. Berger, and W. Dement, A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects, Public Health Service, U.S. Government Printing Office, Washington, DC, USA, 1968.
C. P. Pollak, W. W. Tryon, H. Nagaraja, and R. Dzwonczyk, “How accurately does wrist actigraphy identify the states of sleep and wakefulness?,” Sleep, vol. 24, no. 8, pp. 957–965, 2001.View at: Google Scholar
W. Karlen, C. Mattiussi, and D. Floreano, “Adaptive sleep/wake classification based on cardiorespiratory signals for wearable devices,” in Proceedings of the IEEE on Biomedical Circuits and Systems Conference (BIOCAS '07), pp. 203–206, Montreal, Canada, November 2007.View at: Publisher Site | Google Scholar
W. Karlen, C. Mattiussi, and D. Floreano, “Sleep and wake classification with ECG and respiratory effort signals,” to appear in IEEE Transactions on Biomedical Circuits and Systems.View at: Google Scholar
M. Čepek, M. Šnorek, and V. Chudáček, “ECG signal classification using GAME neural network and its comparison to other classifiers,” in Proceedings of the 18th International Conference on Artificial Neural Networks (ICANN '08), vol. 5163 of Lecture Notes in Computer Science, pp. 768–777, Prague, Czech Republic, September 2008.View at: Publisher Site | Google Scholar
J. R. Koza, Genetic Programming II: Automatic Discovery of Reusable Programs, MIT Press, Cambridge, Mass, USA, 1994.
J. Bongard, “Evolving modular genetic regulatory networks,” in Proceedings of the Congress on Evolutionary Computation (CEC '02), vol. 2, pp. 1872–1877, Honolulu, Hawaii, USA, May 2002.View at: Google Scholar
T. Reil, “Dynamics of gene expression in an artificial genome-implications for biological and artificial ontogeny,” in Proceedings of the 5th European Conference on Artificial Life (ECAL '99), pp. 457–466, Lausanne, Switzerland, September 1999.View at: Google Scholar
T. Reil, “Artificial genomes as models of gene regulation,” in On Growth, Form and Computers, pp. 256–277, Academic Press, London, UK, 2003.View at: Google Scholar
C. Mattiussi, D. Marbach, P. Dürr, and D. Floreano, “The age of analog networks,” AI Magazine, vol. 29, no. 3, pp. 63–76, 2008.View at: Google Scholar
P. Dürr, C. Mattiussi, and D. Floreano, “Neuroevolution with analog genetic encoding,” in Proceedings of the 9th International Conference on Parallel Problem Solving from Nature (PPSN '06), vol. 9, pp. 671–680, Springer, Reykjavik, Iceland, September 2006.View at: Google Scholar
A. Soltoggio, P. Dürr, C. Mattiussi, and D. Floreano, “Evolving neuromodulatory topologies for reinforcement learning-like problems,” in Proceedings of the IEEE Congress on Evolutionary Computation (CEC '07), P. Angeline, M. Michaelewicz, G. Schonauer, X. Yao, and Z. Zalzala, Eds., pp. 2471–2478, IEEE Press, Singapore, September 2007.View at: Publisher Site | Google Scholar
G. Gusfield, Algorithms on Strings, Trees, and Sequences, Cambridge University Press, Cambridge, UK, 1997.
C. Mattiussi, Evolutionary synthesis of analog networks, Ph.D. dissertation, EPFL, Lausanne, Switzerland, 2005.
J. Principe and A. Tome, “Performance and training strategies in feedforward neural networks: an application to sleep scoring,” in Proceedings of the International Joint Conference on Neural Networks (IJCNN '89), vol. 1, pp. 341–346, Washington, DC, USA, June 1989.View at: Google Scholar
D. Nguyen and B. Widrow, “Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights,” in Proceedings of International Joint Conference on Neural Networks (IJCNN '90), pp. 21–26, San Diego, Calif, USA, June 1990.View at: Google Scholar
C. Mattiussi, P. Dürr, and D. Floreano, “Center of mass encoding: a self-adaptive representation with adjustable redundancy for real-valued parameters,” in Proceedings of the 9th Annual Genetic and Evolutionary Computation Conference (GECCO '07), pp. 1304–1311, London, UK, July 2007.View at: Publisher Site | Google Scholar