Outline Why model neural networks? Modeling Neural Networks A - PDF document

Outline � Why model neural networks? Modeling Neural Networks � A brief look at the neuron. � A look at some current works. � Adding an evolutionary strategy. Paul Nuytten CPSC 607 Why Model Neural Networks? The Neuron � The nervous system is a very complex system with many hidden properties. � Many experiments cannot be performed in vivo without destroying the specimen; leaving many questions unanswered. � When looking at a model many of the redundant structures and processes can be removed leaving a more focused picture. � The nervous system is a very efficient and massively parallel computational device. Models may capture this property to solve a certain class of problems. The Neuron The Neuron � The neuron shares many of the same � The unique components are components as many other types of cells. � The dendrites � It’s the unique structures that make the � The axon neuron a powerful communication and � The synapses computational device.

The Dendrites The Axon � The dendrites are short strands that � The axon is a long extension from the protrude from the cell body or the soma. soma. � The dendrites are very receptive to � Each neuron only has connections from other neurons. one axon. � The dendrites carry signals from the � The axon is synapses to the soma. myelinated if it is insulated with Schwann cells. The Axon The Synapse � If the axon is � The synapses are the connections made myelinated the action by an axon to another neuron. potential will travel � When an action potential arrives at a much faster. synapse from the postsynaptic cell, � The axon carries neurotransmitter is released into the action potentials from synaptic cleft. the soma to the synapses. The Synapse What level should be modeled? � The neurotransmitter will interact with ion channels on the membrane of the postsynaptic cell causing them to open letting some ions into the cell while letting other ions escape. � A synapse is call excitatory if it raises the local membrane potential of the post synaptic cell. � Inhibitory if the potential is lowered.

Some Current Works NEURON � NEURON � NEURON is a simulation environment for neurons and neural networks. � GENISIS � The NEURON simulation allows a user to � Neural Swarm focus on the biological and biophysical � Evolutionary Artificial Neural Networks aspects of a neurological system. � Great tool for biologists. NEURON GENISIS � GEneral NEural SImulation System � Provides a simulation environment for biologically realistic models. � Very similar to NEURON. � Allows for parallel processing. Neural Swarm Neural Swarm � Uses the concepts of swarm intelligence when creating neurological models. � Still in its infancy. � Instead of designing the network and defining processes in mathematical terms, the network and processes are allowed to emerge from the simple interactions within the system. � The results are then compared to biologically observed results.

Evolutionary Artificial Neural Evolutionary Artificial Neural Networks Networks � Focuses more on the computational aspect of artificial neural networks (using them to solve problems). � Uses outgrowth and pruning rules to grow a neural network. � Spontaneous neural activity also contributes to the development of the network. � What is nice about this simulation is its ability to apply a genetic algorithm to the above rules and the network’s morphology to specialize the network to a given problem. Evolutionary Artificial Neural Conclusion Networks � When creating a simulation it is important to identify the level at which to model. � It is also important to identify the target audience and intended use of your simulation. � Start simple and gradually add complexity. � Collaborate. References F. Bloom, C. Nelson, A. Lazerson. Brain, Mind and Behavior Third � Edition. Worth Publishers, USA, 2001. N. Campbell. Biology Fourth Edition, pages 993-1009. The � Benjamins/Cummings Publishing Company, Inc., Menlo Park, California, 1996. R. Rojas. Neural Networks A Systematic Introduction. Springer- � Verlag, Berlin, 1996. Rust A.G., Adams R., Schilstra M. and Bolouri H. Evolving � computational neural systems using synthetic developmental mechanisms. 2003. http://www.rwc.uc.edu/koehler/biophys/4d.html � http://www.neuron.yale.edu/neuron/ � http://www.genesis-sim.org/GENESIS/ � http://strc.herts.ac.uk/bio/alistairr/neural_interests.html �