Post by tkorrovi on Dec 22, 2004 19:49:49 GMT -5
From ScienceDaily
www.sciencedaily.com/releases/2004/10/041022104658.htm
"To control the simulated aircraft, the neurons first receive information from the computer about flight conditions: whether the plane is flying straight and level or is tilted to the left or to the right. The neurons then analyze the data and respond by sending signals to the plane's controls. Those signals alter the flight path and new information is sent to the neurons, creating a feedback system."
What I find important is that the neurons seem to learn the same way as the absolutely dynamic systems are supposed to learn
adsproject.sourceforge.net/
ie all we need is to connect the system to the process, no reinforcement signals like in neural networks. They can model and control a process only based on the interaction with the system, something what no artificial neural network can do.
From training the absolutely dynamic systems I at least know how much time it takes for such systems to learn even a simple thing, inevitably a system which has no previous knowledge or control algorithm of a process, has to try many possibilities, which based on my experience takes at least several months in a modern computer. But for 25000 neurons it took 15 minutes. The speed and flexibility of the neurons is amazing, the more we know, the more we probably will be amazed, what a real miracle is a neuron.
Such experiments are extremely important, in order to know how neurons work. But of course a question about ethics would arise, how many neurons there must be to consider such set of neurons an organism, so that it may be considered as a torture of the animals. A nervous system of any animal should not be modified, so such developments should be known in time for people to make up their minds, about where the line goes.
www.sciencedaily.com/releases/2004/10/041022104658.htm
"To control the simulated aircraft, the neurons first receive information from the computer about flight conditions: whether the plane is flying straight and level or is tilted to the left or to the right. The neurons then analyze the data and respond by sending signals to the plane's controls. Those signals alter the flight path and new information is sent to the neurons, creating a feedback system."
What I find important is that the neurons seem to learn the same way as the absolutely dynamic systems are supposed to learn
adsproject.sourceforge.net/
ie all we need is to connect the system to the process, no reinforcement signals like in neural networks. They can model and control a process only based on the interaction with the system, something what no artificial neural network can do.
From training the absolutely dynamic systems I at least know how much time it takes for such systems to learn even a simple thing, inevitably a system which has no previous knowledge or control algorithm of a process, has to try many possibilities, which based on my experience takes at least several months in a modern computer. But for 25000 neurons it took 15 minutes. The speed and flexibility of the neurons is amazing, the more we know, the more we probably will be amazed, what a real miracle is a neuron.
Such experiments are extremely important, in order to know how neurons work. But of course a question about ethics would arise, how many neurons there must be to consider such set of neurons an organism, so that it may be considered as a torture of the animals. A nervous system of any animal should not be modified, so such developments should be known in time for people to make up their minds, about where the line goes.