Paul William Berg

So, the comment spam around here has been insane. I installed a captcha, and hopefully that eliminates the problem for now. I also fixed the Tuktoyaktuk post which wasn't showing up correctly, which I know is a favorite. I also got an automated cross-poster working, so that my blog entries will cross post to live journal automatically.

As I brought up in a previous post, a problem with Koza style genetic programming (GP) is the absence of recursion and any type of variable binding which causes means that the algorithm must unroll all loops and repeat statements because it can't bind statements to variables in order to perform loops. So, in the spirit of Tina Yu, we decide to introduce lambda expression to the set of primitives.

This introduces a new class of problems altogether. The lambda calculus uses binds a variable to a term which is then accessible in the closure, for instance \x->(\y->x+y) is the function f(x) that binds x to the variable "x" and substitutes the value of x with any x in the body, here "\y->x+y". In this case, the body returned is a new function f'(y) that takes a value, binds it to y and substitutes it in the body, "x+y". The body, being a term composed of primitives can now be solved, so that (\x->(\y->x+y) 1 2 ==> (\y->1+y) 2 ==> 1+2 ==> 3.

Ok. for type inference, let me work through this again.

First, lets assume a requirement to return int:

(+ a<int> b<int>)

Now, lets assume we choose the ephemeral constant 1 for a, satisfying the int, and apply (@) for b, with the requirement that the apply return an int:

(+ 1
   (@ c<(-> datum int)>))

The apply must choose a function. All functions are curried to 1 param:

(+ 1
   (@ (lambda (x) d<int>) 
   e<datum>))

Let's now choose for d. we choose if:

(+ 1
   (@ (lambda (x)
        (if f<bool> g<int> h<int))
 

I like Gas Nets. They show great promise. However, there are a couple things I don't like about them, and I think I have devised a solution to these issues.

The traditional Gas Net goes like this:

You have an n-dimensional space, 2 dimensions being the standard, so that's what this example will use. A neuron has a specific X,Y coordinate in this space. The neuron's inputs are defined by 3 numbers. 2 numbers defining the arc extent and another defining the radius, this gives the neuron a semicircle inside which any neurons existing will supply inputs to the neuron. The gas is diffused from a neuron (or from a seperate coordinate) and effects other neurons in the gas's radius. The diffusion is generally modeled loosely, as to not become a fluid dynamics simulation at the expense of processing the neural network itself.

I had visions of using Hindley-Milner style strong typing to constrain the search space of a Koza GP tree. After much struggling with the concept in my head, and not being totally up on my pi calculus, I think I have grokked the impasse. The problem is this:

Say we have a program tree: (+ X Y)

In this case it appears that strong typing is highly adventageous in that it constrains the selection of X and Y to those nodes that yield numbers. A node encoding the string "hello world" would never be chosen for X, as the type inferrence just wouldn't work out. Here's the rub. Let's say we have a "communicate data" function to communicate some unknown data across a network. We can tag this data with a type. But, how do we type the parse tree?

So, if I have a parallel problem, and assign a cluster to it, and my cluster admin server, can add and remove cluster nodes as they become available, then, with a thermostat, I can have the system boot or shutdown computers based on the temperature of the house. This way I have a distributed cluster working on my problem and also working as a standard residential electrical furnace for heating.

Venting to the outside will still need to be performed in the summer when a minimal cluster configuration still generates more heat than is required by the house, but this will be 100% efficient in the winter and as cost efficient as comparable electric heating mechanisms.

Seriously... the whole thing... on an IMAP server no less.

Comeone OSS people... is Kontact seriously the best you can do. My god this thing pisses me off.

I'm going to try evolution again.

I really love scheme's core and I like most of the things in R6RS.

What I really with the scheme community would do, however is completely in the opposite direction:

1) Define 5 or 6 core forms (lambda, define, if, set!, call/cc and maybe 1 or 2 more) that make up the language. This would be what we call scheme and want implementors to provide. Numeric towers and what have you would be part of #2.

2) Define a core library WITH REFERENCE IMPLEMENTATION using only the core forms. Implementors could reimplement this library any way they like for efficiency, but the point is, that if only the core forms in #1 are implemented, the reference implementation would provide the entire language specification.

(link to slashdot)

First, as to the argument that it always takes a more advanced mind to understand a lesser one, that is the very (non-intuitive) thing that the law of universal computation disproves. Given a Turing complete set of primitives, anything operation possible can be performed. A trivial example is simply that since all basic x86 primitives can be decomposed into sequences of the "nand" operation, a simple nand gate can be used to build a C++ compiler, and hence the browser you are reading can be not just decomposed into a series of assembly instructions, but indeed into nand itself, certainly a simpler operation than that of the browser currently being used. The book, Structure and Interpretation of Computer Programs has the introductory student write a full Scheme language interpreter in the Scheme language, then an assembly language register machine in the scheme language, and finally a scheme Interpreter in assembly language, then has the language run itself several layers of abstraction deep just to drill this point in... any universal computer can emulate any other universal computer. Since every law of physics is inately calculatable (even if nondeterministic) then the entire universe can be simulated by a universal computer.

The Chinese room thought experiment only addresses weak AI, hence the distinction. First of all, it assumes that symbolic manipulation without understanding of context is even possible. It has now been quite some time since Turing formulated his famous test and no one has come anywhere close to passing it through context free symbolic manipulation. Personally, I think that such a system is either impossible, or more difficult to construct than a truely intelligent system. If such a thing is possible, then I would not consider it intelligent (and many refute the validitiy of the Turing test on this point). He goes on to refute the "brain simulator" saying that the simulator itself does not "understand" the information flowing through it's synapses. I would also say that the brain doesn't understand it's synapses either. A big assumption in this thought experiment is that the thought experimenter himself is somehow more capable than the experiments he his thinking about, where in reality there is scant evidence that the biological system meets the stringent requirements he is demanding of the artificial system.

The symbolic transform mimicing intelligence view of AI is a very naive view, and one rejected by strong AI researchers. Strong AI does not "act intelligent" via a laundry list of scripted rules. Strong AI has to be taught about the problem through experience and devise a solution to it. In the Chinese room experiment, the system cannot learn based on its experience, therefore it is weak AI, not truely intelligent, and the system, if it one the Turing test tomorrow, would be deemed clever, but inferior to most good AI systems already out there. We have reliable systems today that learn basic language and can function on par with lower mammals. We have not yet created intelligence at the levels that we as humans are at, as they do not ponder their own existence, and we are a long way from doing so. We are not, however, very far from the levels of intelligence we observe in cats and dogs.

To revisit the Chinese room. If the Chinese speaker outside the room described, in chinese, a sequence of rules for performing a novel action (a new dance step for the sake of argument) and the room, after replying with questions to clarify finer points about the process, responded with it's ideas on new flourishes to the steps, then THAT would be indicative of intelligence in the system. I do not think the Chinese room addresses this possibility at all as it assumes a much simpler interaction, perhaps preconcieved bias about the possibility of such interaction interfering with the introduction of such a possibility into the experiment. I do not believe this level of interaction is possible with the describes symbolic manipulation, but I do, strongly believe that such an interaction is possible in true AI, and in fact, recent experiments get these types of responses (on a very simple scale) and rely on such responses to ensure that they are NOT simply doing symbolic transformation.

Quantum Physics is calculatable with math.
Physics is calculatale with math.
Chemistry follows the laws of physics and Quantum Physics.
All elementary brain functions, indeed all physical processes in the universes are describable in terms of one of these disciplines.
A universal computer can calculate all three of these disciplines, and in addition, the discipline of mathematics which can describe an infinite variation of these disciplines not even in existance, AND can describe itself.
A universal computer is trivial to construct. You are looking at one now.
Therefore, everything in the universe can be simulated on a universal computer.
Your brain exists in this universe, or at the very least, exists in concordance with a variation of a chemistry, physics and quantum physics calculatable by mathematics. If you deny this, then the brain is not part of this universe, is spiritual in nature, and cannot be reasoned about logically, and this breaks down into a religious argument.
Therefore, the brain can be simulated on a universal computer, as it is a subset of the universe that can be simulated.

This proof does not take into account the disconnect between variations of universal computers. The Von Neuman machine in front of you may very well be a piss poor platform to simulate the workings of an intelligent mind on and highly inefficient, just as it is horrible at doing quantum calculations and may take decades to do what a quantum computer can do in an instant, but it can be done, and, given that we have working examples of physical minds at our disposal to reverse engineer, I'm sure we can come up with a universal computer suitable for such a simulation given the time.

Finally, none of the above requires a deep understanding of the human mind. The human mind is often studied for ideas because, well, it's the only working model we have to reverse engineer. However, since we are not really interested in human minds (they can be created fairly easily with a couple drinks at your local hangout) in this discussion, but rather, human-level intelligence, we are really interested in modeling the outcome, not the low level process. Most all ideas gleaned from human brain research are rapidly decomposed into their pure mathematical function when implemented into strong AI.

His Response:

Wendy and I left after work on Thrsday, May 25th from Seatac airport, driven there by Seajay. After 14 hours of flying and layovers, we finally arrived in Bangor airport, Maine, and rented a Trailblazer for the trip. We were met at the airport by Chris Anderson, a friend of Wendy's and were escorted, after a brief shopping stop at a local grociery, to his grandmother's house where I was introduced to Jackson and Condin, 2 very adorable women with a very adorable property in Elsworth, Maine. We spent some of the afternoon there and finally made our way out to the coast to Corea, a small lobsterman's village where we would spend the weekend.

We stayed in Chris' cabins on the beach and were met by his Girlfriend, Regina, and another couple, Uve and Sandy and their 2 children. We did much playing on the beach, headed out to Schoodic point for a day, ate wonderful lobster and did a mini road tour of the area. I was attacked by starfish and ate many a mussel collected from the beach there.

There was a sand bar connecting Bar Island to the mainland during low tide and we were able to go exploring out there also. It was a wonderful, relaxing start to the vacation.

Monday morning, we set out along the coast, heading Northeast to Passamaquaddy bay, in search of Pete's Dragon, who we did not see. We did however explore Eastport, an old drinking town with a long history of fishing problems, and coincidentally, the easternmost city in the US. That evening, we passed into Canada, and after a not too brief tangle with customs, entered New Brunswick. With the sun staying up until nearly 10 PM, we traversed New Brunswick mostly in the light and crossed over the Confederate Bridge, the longest bridge over ice covered waters in the world, in the dark. My odometer placed the bridge at around 8 miles long, and it was very impressive.

We decided to stop just south of Charlottetown, Prince Edward Island for the night and have a good sleep in a comfortable hotel. Being just prior to the main tourist season, we had no trouble with heavy traffic or hotels the entire trip.

In the morning, we headed north across the island to the northern coast, where we spent a bit walking the red dunes. PEI's soil is a strange, Martian-like red soil which makes things a bit surreal. The island is mostly made up of small farms and is, for the region, uncharictaristically treeless. Heading West, we reached the fabled Green Gables, home of Anne, a fictional childhood icon of Wendy's. The farm is a real farm and the author spent a lot of time there and accurately depicted it in the story. We toured the rooms of the farmhouse, seeing Anne's bedroom and such, and the barn, and then headed out into the haunted forest, where, along the way, Wendy's childhood fantasy of being kissed on Anne's Lover's Lane was realized, although regrettably, not by Gilbert.

Leaving Green Gables, we went into town to the author's Aunt and Uncle's house (Emily of New Moon's house, for the Green Gable's afficianatos) overlooking the "Lake of Shining Waters" and where the author spent some of her time growing up and where many featured peices of furniture in the books exist in real life, such as the Blue Chest (Which the author locked away private possesions not to be opened until some time after her death) and the enchanted bookcase.

Having been girlie-stuffed out completely, and after a rasberry cordial (served non-alcoholic unfortunately, undoubtedly so that children visiting the premesis can partake in a bright red drink) we headed south again, and back to the bridge, in the daylight this time where we could see it in all it's magnificence.

Back across New Brunswick we drove, this time taking a more northerly route towards Fredrickton. Near Fredrickton, we left the freeway in search of food, only to find we had entered an immense bayou, sans the alligators.

... Long time passes with this in the que and unpublished ...