Can a robot’s software “evolve” to cope with physical changes, such as the addition of new features? That’s the question asked by Robert Gordon University engineer Christopher MacLeod. With the help of colleagues, MacLeod designed a robot that can adapt to changes by “mimicking biological evolution”—or so New Scientist writer Paul Marks claims. Let’s take a closer look.
A robot can adapt to changes by “mimicking biological evolution.”
The team’s robot has a “brain”—a neural network, to be more specific—that attempts to mimic the brain’s learning process. The brain can be trained by giving it priorities, in essence. Marks explains:
For example, if the goal is to remain balanced and the robot receives inputs from sensors that it is tipping over, it will move its limbs in an attempt to right itself. Such actions are shaped by adjusting the importance, or weighting, of the input signals to each node. Certain combinations of these sensor inputs cause the node to fire a signal—to drive a motor, for example. If this action works, the combination is kept. If it fails, and the robot falls over, the robot will make adjustments and try something different next time.
MacLeod’s team has developed a system of artificial “genomes” for the robot to use. An algorithm randomly determines the genome, and the genomes that result in the best robotic performance are “bred” to create better versions.
But the design goes beyond that. When the robot’s structure—its “anatomy,” you might say—changes, the robotic software realizes if it isn’t completing its programmed goal as efficiently and creates new “neurons” to deal with the added features. After trying various combinations, it eventually finds a way to complete a task, at which time the software “freezes” the programming and evolution stops. According to the researchers, the software can even cope with such additions as cameras, “learning” how to seek or avoid light.
So does this lend support for evolutionary theory? In a way, yes, it lends support for the idea that “evolution,” in general, is a valid theoretical concept, useful in various disciplines and contexts. For instance, we can imagine how words and phrases “evolve” within a language according to how popular they are.
None of these are true of supposed Darwinian evolution.
But does the validity of generic “evolution” say anything about whether or not biological evolution occurred, as speculated by Darwin? That is a much trickier question. In instances of societal evolution (words, phrases, fashions, etc.), intelligent designers (viz., humans) are behind the “evolution.” In the case of this robot, while the algorithm is yielding random “genomes,” the evolution algorithm itself—in addition to the entire software system, programming language, and mechanical robot—are all very carefully designed and even given specific goals, with successful evolutions “frozen” at waypoints. None of these are true of supposed Darwinian evolution.
As a case in point, the article reports that the team’s attempt to create an actual independently evolving software “brain” (that is, where the entire brain, not isolated segments, continually evolves) “became too complex and simply ground to a halt.” This underscores how the current system is intelligently managed, and how it bypasses the impossibility of an information-carrying system (i.e., a code or language, such as DNA) arising through chance.
Of course, even if someone could somehow prove that biological evolution (in the molecules-to-man sense) was possible, that’s still not the same thing as proving it happened. What God’s Word clearly teaches is that creation was recent and absolutely supernatural.
For more information:
Remember, if you see a news story that might merit some attention, let us know about it! (Note: if the story originates from the Associated Press, Fox News, MSNBC, the New York Times, or another major national media outlet, we will most likely have already heard about it.) And thanks to all of our readers who have submitted great news tips to us.