Chapter 120 19.9 Uncertain Density of Biological Search Spaces

Darwin was written more than a century ago, before the first computers were invented.Darwin, however, accurately paints a famous picture of evolution in computerized language: evolution "sieves the whole world every day and every hour for the slightest variation; it weeds out the inferior ones." Mutating, retaining and accumulating high-quality mutations; it does all this silently and unknowingly... ".Isn't this the algorithm for searching formal libraries?Is this repository of all possible forms of life a vast space dotted with valid samples, or is it a crowded place?How likely are random evolutionary steps to land on real life somewhere?In this space, to what extent are effective life forms clustered?How far apart are the clusters?

If the possible life-forms were densely populated with viable ones, then natural selection, by sheer luck, would have an easier time searching the space of possibilities.A space full of feasible solutions that can be searched by stochastic methods provides evolution with countless paths that unfold over time.But if viable life forms are sparse and far apart, natural selection alone may not be able to reach new life forms.In the possible biological space, the distribution of surviving life forms may be very sparse, so that most of this space is empty.In this space of failure, survivable life forms may congregate in a small area, or along a few paths that meander through.

If the space of viable organisms is very sparse, then evolution undoubtedly needs some guidance to move from one conglomerate of viable organisms to another across an open wilderness.The trial-and-error approach pursued by natural selection will only quickly get you into a situation where you don't know where you are. We know next to nothing about the true distribution of life in the realistic life-form repertoire.Maybe the distribution is so sparse that there's even only one path to travel through—the one we're currently walking on.Or, several small roads can converge into a broad highway, leading to several necessary passes—for example, attractors with characteristics such as limbs, coelom, and five fingers.Or, somehow, by God's will, no matter where you start, you'll eventually end up with double symmetry, segmented limbs, and intelligence in one way or another.Exactly which is the case, we do not yet know.But if artificial intelligence can make progress, we may know the answer.

People are using a new science—complexity science—rather than biology to raise healthy questions about the fundamental laws of evolution.Much to the annoyance of biologists, the main forces driving post-Darwinian theories have come from mathematicians, physicists, computer scientists, and holists—people who simply don’t work. point.Naturalists have nothing but disdain for those who insist on reducing complex natural phenomena to computer models and have no respect for Darwin, the greatest observer of nature. Speaking of Darwin, he himself reminded readers in the third edition: Recently my conclusions have been misrepresented many times, and it has been suggested that I attribute the changes in species solely to natural selection.Allow me to reiterate that, since the first edition of this book, I have written in the most prominent place—at the end of the introduction—"I am convinced that natural selection is the main, but not the only, process by which species change. way." This sentence obviously did not play its due role.The power of taking words out of context is really powerful!

Neo-Darwinism describes a wonderful story of evolution by natural selection, a story so carefully woven that its logic is almost irrefutable: since natural selection could logically create all species, all species are natural selection Creative.If we can only argue about this in terms of one mode of life on Earth, we have to accept this broad interpretation unless there is irrefutable evidence against it. We have no such evidence so far.All the things I have described here - symbiosis, directional mutation, jumping, self-organization - are far from conclusive.But they do show that evolution has many other factors besides natural selection.Furthermore, a bold and adventurous blueprint is emerging from these questions and fragments-to carry out artificial evolution beyond biology.

When we try to transplant evolution from history to artificial medium, the inner nature of evolution is exposed before our eyes.Artificial evolution running in a computer has passed the first test of neo-Darwinism.It shows that spontaneous self-selection can serve as a means of adaptation as well as generate some primary innovations.