Blogs

In my final project, I want to take several cryptosystems/encryption algorithms and see if it's possible for a neural network to recognize the pattern. If all goes to plan, my dataset will consist of a ciphertext and it's plaintext equivalent; this is what I'd like to train my network with. I'll test simple substitution/transposition ciphers and maybe even ones that rely on the factoring of large primes (ie RSA); however, I highly doubt the network will be able to 'crack' the latter. If time permits, I may also explore the idea of training two networks on one another in an attempt to create an encryption/decryption system. Anyway, I'm running into a problem when I try to set the inputs and outputs for the neural network.

'The line between living organisms and machines has just become a whole lot blurrier. European researchers have developed "neuro-chips" in which living brain cells and silicon circuits are coupled together.' I heard this headline on the way in to school yesterday morning and dug up this link to it. Sounds pretty exciting.. but I don't know enough about neurons to imagine how they would intereact with the electronics. If anyone has an idea of how this might work, could you fill me in?? Also, I thought one of the pictures was interesting.. have a look at the snail neurons.. is that why netlogo calls them turtles??

I saw this article on Slashdot today touting the "first digital simulation of an entire life form," and couldn't help but think of the Karl Sims' "evolved virtual creates" that we saw a few weeks ago during our discussion of evolutionary algorithms. Though the mechanisms and scale of simulation are wholly different (Sims, as I understand it, worked with locomotion mechanisms, while these researchers worked on the cellular level)--the resulting "life forms" existing only in virtual space are quite similar in concept. The premise of both experiments is also remarkably similar in that both groups believed that they could effectively simulate organic life forms using computer algorithms.

I'd like to do a model in NetLogo for the final project, I want to make a simulation of ant colonies based off of Deborah Gordon's work, which I read about in Johnson's Emergence and which was also a part of the talk at the Swarm Exhibit. Anyway, I want to try to model how colony behavior changes as it gets older/larger as well as task allocation based on encounters with other ants/agents and possibly interactions with other colonies. Working on getting some of Gordon's papers now so I'm just starting to play around with ideas how to implement it in NetLogo.

I found a neat article on an unlikely connection between physics and math--a similarity between energy levels of nuclei of heavy elements and Riemann's zeta function. I don't know the first thing about math or physics, but maybe someone with more background can say something interesting about this.

Just posting to say that I have no clue what to do a project about. (Neither does jferraio, she says, though I have great faith that she'll come up with somethine awesome.) I hope I'm not alone. Coding, fine, I can do that. But I just don't know where to start. Anyone have any castoff ideas they're willing to hand down? Jumping points? When I asked Doug about this on Saturday, he asked me what I thought was most interesting in the course so far. I still can't decide. Help?

This isn't particularly good or even described well, but it does reflect my current thoughts on my final project. Perhaps with your critique and commentary, I can refine my thoughts and put together a nice project. Race/Population Relations --A series of simple NetLogo models that illustrate an emergent explanation for complicated social behavior-- Some potential models include: #1. Antagonism between 2 races, with one racial group having superiority in number and resources over the other. - For most individuals of the superior race (red), they would rather live almost entirely with people like themselves (the exact proportion can be on a slider)

If you'll recall, we had a discussion yesterday in class about the "saddle" pattern seen in testing a range of input values (x & y each from 0 to 1) for the xor problem. There was some question as to why the saddle always ran like "/" rather than "\" Basically, the center range of x & y insisted on returning high values, even when Doug added training data for (0.5,0.5) to return 0. My intuition was that it had something to do with the calculations involved in adjusting the network's weights during back-propagation. Math isn't my forte (unfortunately, for a comp sci major!), but I'm pretty sure I've confirmed that the backprop algorithm is biased for answers of 1 over 0. Let me know if there's a hole in my logic, here.

A long time ago, in a galaxy not too far from here, Lindsay Gold brought up the game of "Spore" which is to be released by Electronic Arts in the fourth quarter of this year. I thought that I'd continue that conversation with a new post and a couple of new links. A quick review: Spore is a computer game modeled on evolution, except that there is human interaction. You guide a "spore" (yes, I know that the term is inaccurate) through evolution. As you go along, you can add features to yourself, and your technology progresses as well. You can even acheive space flight! So for those of you interested, here are a couple links:

During Monday's lecture Professor Blank told us about a split into two separate approaches to studying artificial intelligence. The split is between the rational models and emergent models. Professor Blank writes that "These two paradigms, in my opinion, have little to do with one another. That is, emergent models can certainly show rational, rule-like behavior. But the implementation of emergent models have nothing to do with how rational models operate." I don't understand how this dichotomy is possible! I agree that emergent models can show logical "rational" behavior. We need only look to Langton's Ant and look at its rule like behavior when building its road to believe it.