Can Evolution Be Programmed?
Some researchers are employing “evolutionary computing” as an algorithm to solve problems. But is it really evolution?
- Evolved machines: A company called Evolved Machines in Palo Alto announced a 40-teraflop machine that will be used for the “artificial evolution of neural circuitry” (see press release on United Business Media). “It is self-evident that in biological brains exquisitely complex neuronal circuits wire themselves together,” The Evolved Machines website says. “Further, neuroscience research has recently established that these neural circuits continue to rewire themselves during life, embedding information about the outside world and internal activity alike.” Examples are the brain, the olfactory organ, and the eye. OK, so what does this have to do with evolution? The press release states, “Simulated evolution can be used to guide the selection and parameterization of these mechanisms in simulations of highly neural circuit fabrics, provided an enormous amount of parallel computing power can be applied.” They call this “reverse-engineering circuitry in the brain to enable a new class of self-wiring devices that perform in the complexities of real-world environments, for both artificial olfaction and visual object recognition.”
- Selecting natural laws: Can a computer running an evolutionary algorithm play Isaac Newton? That’s what an article on Science Daily suggests: Evolution is helping Cornell scientists discover natural laws. “The researchers have taught a computer to find regularities in the natural world that become established laws – yet without any prior scientific knowledge on the part of the computer.” The Cornell researchers trained their algorithm to look for “invariants” while computing derivatives of every variable in a system. “Then the computer creates equations at random using various constants and variables from the data,” the article explains. “It tests these against the known derivatives, keeps the equations that come closest to predicting correctly, modifies them at random and tests again, repeating until it literally evolves a set of equations that accurately describe the behavior of the real system.”
But can this really be called evolution? “All equations regarding a system must fit into and satisfy the invariants,” said Michael Schmidt, a specialist in computational biology. “But of course we still need a human interpreter to take this step.” Some other “cheating” was involved:
The researchers point out that the computer evolves these laws without any prior knowledge of physics, kinematics or geometry. But evolution takes time. On a parallel computer with 32 processors, simple linear motion could be analyzed in a few minutes, but the complex double pendulum required 30 to 40 hours of computation. The researchers found that seeding the complex pendulum problem with terms from equations for the simple pendulum cut processing time to seven or eight hours.
This “bootstrapping,” they said, is similar to the way human scientists build on previous work.
Can this be compared to what biology does, or did? The researchers said the computer takes care of the grunt work, “helping scientists focus quickly on the interesting phenomena and interpret their meaning.”
- Evolving war: French scientists got a virus and a bacterium to undergo a co-evolutionary arms race, reported Science Daily. By running some “experimental evolution” using Darwinian selection, they watched the predator and prey evolve to outwit each other. The evolution, however, seemed limited to whether the bacteria formed a biofilm or sat at the bottom of the bottle. Both forms may already have been present. It seems that one form or other was resistant depending on the conditions under which the predator virus was added to the mix. Either way, it was just a game of last bacterium standing, without knowledge of how they succeeded. They said, “What makes prey resistant or predators capable of attacking them again remains poorly understood.”
Speaking of biological computation, Live Science wrote up something for baseball fans: “How Baseball Players Catch Fly Balls.” Apparently good players know how to gauge the vertical acceleration of the ball to determine whether to run toward the ball or away from it. Counter-intuitively, almost all players start by running toward it. The reason may be to accentuate the measurement of vertical acceleration. “A faster rise of the optical acceleration above the detection threshold may outweigh a possible initial step in the wrong direction,” the article explained. “Making an initial step forwards is not only easier than making an initial step backwards, but might also be a better choice.” Coaches should be patient with Little Leaguers, the article ended, saying that “Their brains may still be learning the math.”
Amazing as some of the research results are, this entry gets the Dumb category for assuming this is like evolution. Anything that involves intelligent selection of outcomes is as far from Darwin as an earthquake from city planning. Material particles do not understand and interpret natural laws, nor do they build systems. The equivocation of the word “evolution” in these intelligently-designed research programs with what Darwinists are talking about is perverse. It amounts to a snow job, stealing glory for Charlie from ID projects. Darwin gets no more credit for these interesting results than Kim Jong Il for inventing democracy. Progress in the creation-evolution debate can only be made by everyone agreeing to definitions and terms and rules of argument. Researchers, get your purposeful hands off the apparatus. Care nothing about what happens. Don’t select outcomes or interfere in any way. Then, as everything collapses in a heap of entropy, you will begin to understand the resources available to the kind of evolution Darwin preached. For a taste of common sense to melt the snow job, read this article by The Country Shrink. Notice especially the quote by D. L. Abel.