Hi everybody I am wondering what effect the error parameter for each individual has. I am currently working on heads-up poker players. Each player is assigned...
This is all very interesting to me as I hadn't really thought about the single value approach. I think my mind briefly forayed into the possibility of a ...
John, thanks for the detailed response..It will take some time for me to digest... not initializing the neurons in the net seems strange to me. if you have a...
John, I want to add that the only difference between what I was doing and what you propose was the resetting of the net. Of course I only calc'd dependencies...
... A wise move... :) ... What is 'fair'? REAL networks don't have the privilege of being able to see the OLD state of the neuron they are connected to... ...
... Yes, outputs are inputs. It's all a big loop. But it's more than just that... When the 'expected' input (which is an output of an 'expectation' circuit)...
... Oh, I see what you are doing... Yes, each time you calculate the node, you calculate the sum of its inputs, apply the sigmoid, and set its output value....
... Well, NEAT doesn't address this specifically but I'm not sure what you're saying here. It's certainly "handled" by NEAT as well as anything else ...
John I have to agree with Jeff here...my instincts tell me that NEAT should be highly amenable to finding optimum network structures that have desired...
... Oh, sorry, I made a cognitive leap that I didn't let you in on. My bad... The assumption that I had was that the topology of the network was constantly...
... I remember that story, too. That is an oversimplification. The memory isn't stored in that one neuron. But that neuron always gets activated whenever...
... I guess it all comes down to this and I disagree. I imagine that when you learn, it is not a topological innovation. I think that would make for an ...
... Really? Then how do you propose that learning can take place in a fixed topology, fixed weight network? Sure, with a properly recurrent network, previous...
... Going back to the discussion about spiking neurons and what info might be in the spike trains. One idea I had was that the spike trains convey addressing...
... Or both. I would imagine you can fix both and still create a learning machine. But I think Colin summed it up nicely. I think you're sort of stuck on...
Colin – After working through most of the forum messages and considering various NEAT environments to work…okay, play…with, I downloaded your SharpNEAT....
... That's an oversimplification. The image of a horse is hundreds of thousands of individual signals entering the brain. Those signals enter the nural...
Well, I'm sure people are getting weary of hearing us go on about this particular topic so I'll just say one last thing on it. - One can recreate any arbitrary...
... Interesting. I'd like to see that borne out. It would be fantastic to see a neural network that can retain information without being subject to losing it...
I'm not sure I follow this logic. A computer isn't just made out of NAND gates. It also has flip-flops for storing data. Therefore, I don't think it follows...
Lars, I believe the error variable is just there for convenience. That is, I don't think NEAT actually uses that number to make any decisions. The field can...
Ken, I don't know if Jeff is calling a ANN with recurrent connections, static or not. If he means the network is static topologically, then he could be...
... I would assume so since Jeff was making the point that NAND gates can be used to build any logic circuit and these circuits (in CPU's etc) often utilise...
... Here's a link to details for a 'D' (for Data or Delay) flip-flop, note it uses only NAND gates... http://www.play-hookey.com/digital/d_nand_latch.html ...
On Mon, 05 Sep 2005 13:22:15 -0500, Jim O'Flaherty, Jr. wrote ... Yes, and I have done a very poor job of saying as much. While I think that feed-forward...
Hello everybody I started dabbling in AI abouit twenty years ago on and off. For a while I thought that genetic programming could be an avenue to follow ...
John (caloyannis) is right when he says the "trick is to find it!" I think it's important to keep in mind in discussions such as this one that theoretical...
At the risk of offending the gods, I'm going to humbly add 2 cents. While I fully agree with everything you've said here, I think there is an important...
Jeff, I believe we are in agreement. I just wanted to make sure we don't stop thinking about the interesting practical issues. The theoretical side deserves...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
