Ns-2.33 (and nam) on Ubuntu 8.04 (Hardy Heron)
ANN for Wireless Network Applications
September 28th, 2009 by Jason Ernst
This semester I have been taking a soft computing course. We have covered fuzzy logic and are starting artificial neural networks (ANN) although I have missed a couple of classes due to the conferences I have been attending. Anyway the ANN class today piqued my interest in how I can apply this to my area which is wireless networks. It seems to me so far that it could be applied to some of my cross-layer work since the network could be trained to tune parameters to settings which yield good performance based on specific network conditions. However, I’m not sure if this approach would be good or if some other AI type of technique may be better. Also I am interested in how ANNs could be applied to breaking encryption schemes if it is even possible. I have tried a few searches on Google and some journals / conferences but nothing of interest has come up yet. I don’t think I really understand ANNs enough to answer any of the questions, but I thought I’d get them down so I can come back later and think when I have more time. Feel free to leave any comments or suggestions on these ideas.
Posted in Computer Science, Miscellaneous, Research, Wireless Networks by Jason Ernst
Tags: Artificial Neural Networks, Encryption, Security, Wireless
October 12th, 2009 at 7:22 pm
Hey! Are you taking Stefan's soft computing course?
Oddly, I only ever worked with gradient descent neural networks– I only really started looking into more interesting training patterns after I left (and have yet to have time to truly implement or compile others' code)…
If you have time, see if you can make heads or tails with something called "BGFS".
October 20th, 2009 at 8:21 pm
Hey, yes it is Stefan's soft comp course. I will try to take a look at this BGFS you speak of. It's too bad I didn't see it before I did my survey paper, I might have been able to incorporate it into the paper