Chapter 109 18.3 The difference between learning and evolution

The exciting links between learning, behaviour, adaptation and evolution have only been studied in recent years.The vast majority of the work is carried out through computer simulations.Biologists used to look down on these jobs more or less—but not anymore.Simulations by researchers such as David Ackley and Michael Littman (1990), Jeffrey Hinton and Steven Nolan (1987) have unambiguously revealed that the How populations of organisms that learn evolve faster than populations of organisms that do not learn.Learning here refers to the continuous search for possible adaptations by changing behavior.In the words of Akeley and Littman, "We found that organisms that can combine learning and evolution are more successful than those that only learn or only evolve, producing more resilient populations that are consistently Survive until the end of the simulation." In their simulations, the heuristic learning performed by the organisms was essentially a random search algorithm for a definite problem.At the first European Artificial Life Conference held in December 1991, the experimental results submitted by two other researchers, Parisi and Nolf, showed that the self-directed learning of tasks selected by the biological group has the best learning efficiency. , and the adaptability of organisms has thus been strengthened.They boldly assert that both behavior and learning are among the drivers of genetic evolution.This assertion will increasingly be accepted by biology.

Going a step further, Hinton and Nolan speculate that Baldwin's theory is most likely to apply to problems that are particularly "rugged."They argue that "for biologists who believe in regular topographical ups and downs in evolutionary space ... the Baldwin effect makes little sense, and for biologists who question the well-structured nature of the search space, The Baldwin effect is an important mechanism that allows organisms to use the adaptive processes in their bodies to greatly improve their evolutionary space." Organisms create their own possibilities. Michael Littman told me, "The problem with Darwinian evolution is that you have to have enough time to evolve!" But who can wait a million years?Among the efforts to infuse artificial evolution into manufacturing systems, one way to speed things up is to add learning to them.Artificial evolution will likely require a certain amount of artificial learning and artificial intelligence to be staged on a human-acceptable timescale.

Learning plus evolution is a recipe for culture.The transmission of information to genes through learning and behavior is genetic assimilation; conversely, the transmission of information from genes to learning and behavior is cultural assimilation. Human history is a process of cultural inheritance.With the development of society, human learning and imparting skills echoes the memory and ability inherited in the biological sense. From this point of view-and this idea is actually a long-standing one-every cultural advancement (slash, fire, writing) achieved by previous humans has prepared "possible space" for the transformation of human mind and body , so that the biological behavior of the past can be transformed into the cultural behavior of the future.Over time, as culture took on some of the biological work, human biological behavior gradually became dependent on human cultural behavior and more effectively supported the further development of culture.Every extra bit that children pass on from culture (the wisdom of their ancestors) rather than animal instincts gives biological humans an extra chance to pass that culture on from generation to generation.

Cultural anthropologist Clifford Geertz sums it up: "The slow and firm cultural growth throughout the Ice Age, like a glacier, changed the selection pressure faced by evolving humans and played a major guiding role in human evolution. Although the details are difficult to trace back, the perfection of tools has Organized hunting and gathering, true family structures, the discovery of fire, and, more importantly, the growing reliance on symbolic systems (language, art, myth, ritual) for communication and self-restraint, all contribute to the Created new environments to which we had to adapt... We had to abandon the path of trying to control our behavior precisely and regularly through our genes..."

If we think of culture as a self-organizing system—a system with its own agendas and pressures to survive—then human history becomes more interesting.Richard Dawkins has stated that those self-replicating thought or meme systems can quickly accumulate their own affairs and behaviors.I think that for a culture, the most basic drive is to reproduce itself and change the environment to facilitate its spread, and nothing else.Consuming human biological resources is one way that culture, a self-organizing system, can survive.And human beings are often good at some specific jobs.Books free the human mind from the long-term burden of storing information so it can do something else; language compresses clumsy gesture communication into time-saving and labor-saving sounds.Through the change of generations, culture will carry more and more functions of organisms.Sociobiologists Wilson and Charles Ramsden used mathematical models to discover what they called the "millennial rule."Calculations show that cultural evolution can drive major genetic changes that catch up within a millennium.They reason that the cultural upheavals humans have undergone over the past millennium may have some genetic undertones—although changes at the genetic level may be invisible to us.

Wilson and Ramsden argue that genes and culture are so tightly coupled that "genes and culture are inseparably linked. A change in either will inevitably force a change in the other." Cultural evolution can Shape the genome, but it can also be said that genes also have an inevitable impact on culture.Wilson believed that genetic change was a prerequisite for cultural evolution.If genes aren't flexible enough to adapt to cultural changes, they won't be able to take root in the culture for long. Culture evolves with our bodies, and vice versa.Without culture, human beings lose their unique gifts. (A less pertinent piece of evidence is our inability to turn animal-raised “wolf babies” into creative adults.) Culture and physicality merge into a symbiotic relationship.In Danny Hillis's conception, civilized humans are "the world's most successful symbionts"—cultural and biological behaviors that are mutually beneficial and interdependent—and are a most brilliant example of coevolution.Like all coevolution, it follows the laws of positive feedback and increasing returns.

Culture reshaped organisms (or rather, allowed organisms to reshape themselves) to make them suitable for further cultural development.Therefore, culture tends to be a process of self-acceleration.Just as life breeds more numbers and types of life, culture also breeds more numbers and types of cultures.What I mean here is an enhanced process: an organism guided by culture is biologically better suited to produce, learn, and adapt, and in culturally rather than biologically.This means that we have brains that create culture because culture requires and produces such brains.That is to say, no matter how tiny cultural fragments the species before the emergence of human beings had, they would be helpful for the successors to create more cultures.

For the human body, this accelerated evolution towards information systems seems to imply biological atrophy.From the perspective of learning and knowledge accumulation, culture is a self-organizing behavior that expands itself at the cost of biology.Just as life ruthlessly invades matter and appropriates it, so culture appropriates biologicality.Here I claim that culture modifies our genes. I don't have the slightest biological proof for this.I've heard a statement from Steven Jay Gould et al that "Humans haven't changed in any form since the Cro-Magnon 25,000 years ago." I don't know this though What the claim means for my claims, and not sure how accurate Gould et al's claims really are.On the other hand, biological degradation is astoundingly rapid.The lizards and mice that inhabit the dark caves are said to lose their vision at any time.It seems to me that the body will throw off some of its daily drudgery whenever it gets the chance.

What I'm trying to say is that the advantages of Lamarckian evolution are so significant that nature found a way to make it happen.In Darwinian terms, I would describe its success in this way: Evolution is constantly scrutinizing the world, not just to find better organisms, but to find ways to improve its abilities.Every moment it seeks to make some progress in adaptation.This relentless self-push creates a tremendous pressure—like an entire ocean looking for a leak through—to improve its adaptability.Evolution scours the surface of the planet, looking for ways to speed itself up, to make itself more flexible, more "evolutionary" - not out of its own will, but because ever-accelerating adaptation is out of control The track, on which it drives, cannot help itself.It is searching, trying to find the Lamarckian evolution without knowing it; the Lamarckian evolution is just the gap that is smoother and more evolvable.

As animals evolved complex behaviors, evolution began to break free from the Darwinian shackles.Animals can respond to external stimuli, make choices, migrate to new environments and adapt to environmental changes, all of which have created conditions for quasi-Lamarckian evolution.And with the evolution of human brain, human beings created culture, and culture gave birth to Lamarckian evolution with real acquired inheritance. Darwinian evolution as a learning process is not only slow - in Marvin Minsky's words - but "dull".When the most primitive brain tissue was born, evolution discovered that introducing active learning could speed up the pace of evolution.And when the human brain was born, evolution finally found the complexity it needed to foresee and direct its own course.