Never Ending Math Equation

I told myself that I would post everyday, and I missed a day. Shame on me... in fairness, I had exams et al yesterday.

So the semester is officially over for me, and I am looking forward to next semester. I am taking Neural Networks and Natural Language Processing, which are both Artificial Intelligence classes. Just to dork it up for a minute--

(...thinking how bad it is that this entire post will be dorky...)

One of the interesting things about computer science is its intersection with humans. The human brain is an absolutely amazing processing machine. Computers today work very much like the brain but on a much much smaller scale. The brain exploits parallelism, meaning that it can do many many things at once. Your brain is like a lot of smaller computers put together. However, there are theoretical things in computer science that allow us to make statements that apply to humans.

For example, in computer science there is a concept of a 'Turing Machine'. This is a model of computation that describes a generic machine for solving problems. I won't go into all of the gory details, but basically a Turing machine is the canonical model for modern computers.

The question is "Is the human brain a Turing Machine?". Due to theoretical work in this part of computers science, if the brain IS a Turing Machine then that means that it is equally powerful to every other Turing Machine (of which computers are one). What this means is that computers can one day be indistinguishable from humans. Let me make this clear--- if the brain and the modern computer are the same model of computation, then that means that computers fundamentally must be able to solve the same problems as the brain. This is interesting, because it means that there is not a theoretical limitation on how powerful computers can be--its just an issue of scale.

If computers are as powerful as the human mind, then one day we can interact with computers, and we would not be able to distinguish between interacting with a computer or interacting with a human (this is the prized Turing Test). This is profound to me. There is something in AI called "The singularity" and it is a hypothesis that states that there will be some point at which humans will create a computer smarter than himself. Once this occurs, there will be a massive explosion of technology. This is because, as soon as you build one computer smarter than yourself, then it can build one even smarter and so on and so on. (I will post more on my thoughts on this later in another dork-fest post.)

Lastly, if the human brain is NOT a Turing Machine, then it must differ in some exotic way. Small variants of Turing Machines can all be reduced to the standard turing machine (this is hard to explain without details). But basically there would have to be some property of the brain that is exotic and fundamentally different from the Turing Machine model of computation. For example, if you have watched "what the bleep do we know" then you have seen the theory of how humans must be time traveling small periods of time ALL THE TIME in order just to do everyday things like "hit a baseball" (they invoke the all-holy quantum physics explanation for everything that doesn't make sense and make it really bizarre and exotic). The claim is that the human brain isn't smart enough to do all of the processing necessary to actually hit a baseball. This really isn't true-- they assume that the brain is doing all of the precise mathematics (calculus) to determine how to position the bat. This is false-- the brain uses simple heuristics all the time to do these things. If the brain did calculus subconsciously, then it wouldn't get things wrong so often. The fact that batters strike out is a good hint that many heuristics (which aren't perfect) are being employed.

All of this is very interesting to me...I am waiting for the machines to take over and make me bionic so that I can live forever :-)

Much Love and sorry for the dork-tacular post...hehe

Steve