A local optimum is not a global optimum

Reyna 2022-03-19 09:01:03

The original link https://soulogic.com/item/2833 is abridged Last night I watched an old sci-fi movie Gattaca from more than 20 years ago, which ostensibly raised a question about genetic discrimination. That's really an old movie. Three years later, in 2000, mankind announced that the working draft of the Human Genome Project was completed. But the realism of the absurdity reminded me a lot of this idea.

In addition, the topic of genes reminds me of Metal Gear Snake . The protagonist and the villain are also biological brothers. The villain hates that he is an inferior gene and the protagonist inherits the good gene, but in fact the opposite: the villain is the good gene. , while the protagonist's inferior genes win. I think this line goes well with Gattaca too

The most you can say about DNA is that it governs a persons potential strength, potential destiny. You mustn't allow yourself to be chained to fate, to be ruled by your genes. Humans can choose the type of life they want to live.

When I saw Wang Jianshuo talking about General Electric ten years ago , I was surprised that there are still companies in the world that are forced to be eliminated from the bottom (although I heard that this practice has been abandoned now). At the time, I vaguely felt that this was wrong, but Can't say what's wrong.

Later, it can be explained with a slightly more accurate word that the local optimum is not the global optimum, as explained in the greedy algorithm on the wiki

Greedy algorithms are particularly effective in problems with optimal substructure. The optimal substructure means that the local optimal solution can determine the global optimal solution. Simply put, the problem can be solved by decomposing it into sub-problems, and the optimal solutions of the sub-problems can be recursive to the optimal solution of the final problem.
For most problems, greedy methods usually fail to find the best solution (though there are exceptions) because they generally do not test all possible solutions. Greedy methods tend to make decisions prematurely and thus fail to reach the optimal solution.

The problem with GE is that it simulates natural selection pressure with forced end culling without preserving enough species diversity. Of course, this is only one of the reasons. There are many practical reasons. For example, human beings will of course be smart enough to take advantage of KPIs.

This leads to a new question, R/K strategy , the word is seen on Zhihu: Is Social Darwinism wrong? why?

Watching Westworld a few years ago , I was amazed by one of the famous quotes:

"Mistakes" is the word you're too embarrassed to use. You ought not to be. You're a product of a trillion of them. Evolution forged the entirety of sentient life on this planet using only one tool: the mistake.

I've learned from my work that all "features" are hard to categorize as absolute advantages or disadvantages. Let's say someone has a very good temper and hardly ever confronts anyone, and I've always thought that my often undisguised arrogance was a weakness. But when I took over as the head of the group, I thought he was the logical choice, but different people (even those who were usually very shrewd) told me frankly that he was too weak. I later realized that my combination of traits made me who I am, and even some of the traits I always thought were strengths were actually made of weaknesses. I will write another article on this later. Let’s take a simple example. I always think that my research ability is very strong, but in fact, my memory is very poor, and things that I am not interested in will soon be forgotten, such as remembering people’s names. very difficult. This allows me to passively and effectively eliminate distractions, and as a result, my overall memory and focus are weak, but instead it allows me to use my memory and focus more efficiently, resulting in better output .

At that time, I also thought about a question, if human technology is advanced enough to be like God, a planet with a similar environment to the early earth, and an environment with a large number of creatures, put carefully designed species to compete with it, I am very willing to bet, The dominant species this week in 100 million years will not be the designed ones, and even if the designed species hadn't gone extinct, the evolutionary gap would have been no greater than the gap between dinosaur-era mice and modern-day dolphins. small. The theory of design will not work, nor will it work if God comes. The reason is that the resources of the universe itself are not enough to perform calculations on the universe, and our best program is only a human who has just played the game of Go. Ken Thompson, one of the grand masters of computer, once created a database of chess endgames . Only after exhaustive exhaustion can we say with confidence that when there are the last few pieces left in chess, we can know who will definitely win, which follows the first piece. Go is too far away, and Go, as a highly simplified version, is too far from nature in reality (reality also includes the "table flipping" of asteroid impacts).

View more about Gattaca reviews

Extended Reading

Gattaca quotes

  • [last lines]

    Vincent: [voiceover] For someone who was never meant for this world, I must confess I'm suddenly having a hard time leaving it. Of course, they say every atom in our bodies was once part of a star. Maybe I'm not leaving... maybe I'm going home.

  • Vincent: Is the only way you can succeed is to see me fail?