Chapter 42 edge of chaos

edge of chaos Farmer said that although connectionist models are promising, they are far from revealing the full mysteries of the new second law.First, they fail to describe how emergent phenomena arise in economic, social domains, or ecosystems in which "nodes" are both intelligent and mutually adaptable.To understand such a system, one must understand cooperation and competition in the dance of co-evolution.This means doing research with models of coevolution, such as Holland's ecosystem model, which has become increasingly popular in recent years. What's more, neither the connectionist model nor the coevolutionary model reveals the fundamental question of why life and mind arose.What about a universe capable of generating life and mind?Just talking about "emergence" is not enough.The entire universe is full of emergent structures, such as galaxies, clouds, and snowflakes, which are merely physical objects without any independent life to speak of.There must be another reason for this.And this hypothetical new second law will tell us why.

Clearly, this work relies on computer models that seek to understand the world of fundamental physics and chemistry, such as the molecular automata models that Langton was passionate about.The strange phase transitions at the edge of chaos that Langton discovered in molecular automata appear to provide a large part of the answer, Farmer said.At the Artificial Life Symposium, Langton was careful not to say anything about the subject because he hadn't finished his doctoral dissertation, but many at Los Alamos and Santa Fe found the concept of the edge of chaos fascinating from the start.What Langton was basically saying was that this mysterious "thing" that gave rise to life and mind was some kind of balance between the forces of order and the forces of disorder.More precisely, Langton was saying that you should be looking at how a system works, not what it's made of.When you look at a system from this angle, he said, there are two extremes of order and chaos.It's very similar to the difference between a solid, where the atoms are locked in place, and a fluid, where the atoms tumble around each other at will.But right in the middle of these two poles, in a phase of some sort of phase transition abstractly called "the edge of chaos," you find complication: behavior at this level, where the elements of the system are never quite locked in place. , but never disintegrated to the point of riot.Such a system is both stable enough to store information and fast enough to transfer it.Such systems are autonomous and adaptive living systems that organize complex computations in response to the world.

Of course, strictly speaking, Langton only demonstrated the relationship between complexity and phase transitions in the molecular automata model.No one really knows whether this can be used to explain other computer models, or to explain the real world, too.But on the other hand, there are indications that Langton's discovery may have general significance.For example, after the fact, you will find that in these years, half of the models of correlation theory will have behaviors similar to phase transitions.Back in the sixties, one of the first things Kaufman discovered in his gene networks was phase transitions: if the points of association are too sparse, the entire network essentially freezes and stands still; if the points of association are too dense, , the entire network will churn violently, showing a state of complete chaos.Only between the two, when each node has only two inputs, can the entire network produce the kind of stable cycle that Kaufman wanted.

By the mid-eighties, Farmer said, the same was happening with the autocatalytic group model.This model has many parameters, such as the catalytic strength of various reactions and the supply rate of "food" molecules.Farmer, Packard, and Kaufman had to manually tune these parameters through trial and error.One of the earliest things they found in the autocatalytic group model was that until these parameters fell into a certain range, the autocatalytic group would not kick in and develop rapidly.This behavior is a reproduction of phase transitions in other models, Farmer said. "We feel the similarity, but it's hard to define it precisely. This is an area where someone needs to do careful comparisons, to establish some kind of general framework that was described in the Rosetta Stone paper. "

At the same time, there is even murkier understanding of whether this notion of the fringe of chaos also applies to coevolutionary systems.In ecological or economic systems, Farmer says, we have no idea how to define exactly concepts like order, chaos and complexity, let alone phase transitions between them.But even so, the law of the edge of chaos always makes people feel that there is some truth.Taking the former Soviet Union as an example, Farmer said: "It is now obvious that using centralized methods to control society will not have good results." The control is too tight, so it cannot be maintained.Or you can take the example of the Big Three in Detroit in the 1970s. These car companies have grown too large and too rigidly locked into a certain way of operating, so it is difficult to recognize the challenge from Japan. Increasingly, it becomes even more overwhelming to respond to this challenge.

On the other hand, anarchism is not an effective social mechanism.Some parts of the former Soviet Union seem to have demonstrated this after the collapse of the Soviet Union.A laissez-faire social system will not work.The industrial revolution in England in Dickensian horror novels, or the more modern collapse of savings and loans in the United States, illustrate this point.It is common sense, not to mention the revelation of recent political experience: that any healthy economy and any healthy society must maintain a balance between order and chaos, not some impotent, mediocre, in-between Balanced like a road.Like living cells, they must adjust themselves in a web of feedback and control, but at the same time allow ample leeway for creation, change, and feedback to new situations."Evolution thrives in a resilient system organized from the bottom up," Farmer said. "But at the same time, in that system, bottom-up activity has to be channeled so that it cannot destroy the organization." Evolution is only possible if there is structure.” The complex dynamics on the edge of chaos seem to be the ideal explanation for this evolutionary behavior.