- Share
-
-
arroba
As some UD readers are aware, one of my interests is artificial-intelligence computer programming, especially games-playing AI (here, here, and here).
In producing retrograde endgame databases for the game of checkers, with massive computational resources (two CPUs performing approximately a billion integer operations each per second over a period of two months, for a total of 10,000,000,000,000,000 [ten thousand trillion] mathematical calculations), some very interesting results were produced, including correction of human play that had been in the books for centuries. But did the program produce any new information? Well, yes, in a sense, because the computer found stuff that no human had ever found. But here’s the real question, which those of us at the Evolutionary Informatics Lab are attempting to address: Was the “new information” supplied by the programmer and his intelligently designed computational algorithm, or did the computer really do anything original on its own, in terms of information generation?
The answer is that computers do not generate new information; they only reshuffle it and make it more easily accessible. Here’s an example:
Assume that a census has been taken, and we have data (information) about the annual income of every individual in the world — about 6.6 billion people. Suppose that we would like to know the average (mean) income. A computer program, given the data, could generate this new “information” almost instantaneously, but it would be impossible for humans to do this calculation with pencil and paper because it would simply take too long, and the probability of error would be too high.
Would any new information be generated by such a computer program? It might appear that the answer to this question is yes, but the real answer is no; all the information was included at the beginning, in the data and the computational algorithm. The computer program only made the information more easily accessible and understandable in a reasonable amount of time.
I’ve coined a phrase for this phenomenon. I call it “the elegance of brute force.”
The relevance of this dissertation to the origin of biological information should be obvious. Can random variation (mutational or of any other stochastic variety), filtered by natural selection, produce novel information, especially the information that created humans who design computer programs and create real novel information?