I had started several courses on coursera.org related to neurobiology or computational neuroscience. They are getting more and more complicated yet more and more interesting.
Today I finished another lecture from “Synapses, Neurons and Brains” that discussed the Wilfrid Rall’s Cable Theory and Dendritic Computations. After so many week studying how the neurons work, how they pass electrical signals from axons to dendrites, from spikes to post-synaptic potentials, I started to dive into the computational aspect of the brain.
This lesson discussed several aspects of the brain computation, first at the level of a single neuron, and then, at the level of electric-distributed trees; from the theories that support today’s computational models to complex breakthroughs that develop our understanding of the neural network.
Recognition algorithms have been implemented with artificial neural networks. I have been intrigued by them as well as genetic algorithms; I have thought of the concept Genetic Neural Networks. Now, I am presented with electrical-distributed dendrites and I am starting to think about distributed neural networks – if one network takes care of the auditory system (the audio input in a computer), another may take care of the visual system (the visual input from a camera). The question is: would a genetic neural network be able to change itself to adapt sub unit networks to perform and communicate better with mutation in the blueprints of the network itself? Can we create a neural network with self-awareness?