This article from the micro-channel public number: Eden peaches (ID: PeachAtEden), of: fujiaa, FIG head Source: UNsplash
This article is a study note for "Algorithms to live by".
More than two years ago, I bought a house in the Bay area of California. The Bay area housing market is not comparable to the big domestic cities, but programmers are fierce after all. The average time of each house in the market is about 4 to 5 days, can receive more than 20 bids, the average high price gets. As a mean person, I am always worried about making a huge decision to invest in a large sum of savings. Every time you see a house, you don't stop thinking: is there anything better? Will there be more bathrooms? Is there anything closer to the subway station? Is there anything cheaper? The house was gone before we had finished discussing it with Mark.
In the book "Algorithms to live by" written by Tom Griffths, a professor of psychology and cognitive science at the University of California, Berkeley, and Brian Christian, a best-selling author, this situation is attributed to the famous brainstorming topic "Secretary Question" in the field of mathematics: Imagination You are hiring a secretary, you want to find the best person among the candidates, but you know nothing about the candidate, you don't know what kind of person will come to apply, so you can only take turns to interview each person. . Although you can always accept any candidate, if you don't respond to the other person immediately after the interview, he will disappear and go to work elsewhere. How can you guarantee that the candidate you are interviewing is the best choice?
The omnipotent algorithm gives the answer: 37%. You need to set a total recruitment time for yourself. The first 37% of the time is only for interviews. Collecting data is not a decision. After that, if you meet someone who is better than all the previous candidates, you can get the most notice immediately. You can get the most. The probability of a good candidate is 37%. This is the best you can get.
It sounds like a 37% success rate is a very low number. But this algorithm can be applied to any big data. If you have 100 candidates, the chances of getting the best employees are actually only 1%; and if you have 1 million candidates, your chances are only one in a million. In this case, the 37% success rate is already a very good number. Of course, in order to achieve this success rate, you also need to pay the time and energy for collecting the previous 37% of the data.
When I read this book, I immediately called Mark, who was skiing in Lake Tahoe, and asked him to go home quickly to see the house. We only have 6 months to buy a house, assuming that the number of houses on the market is evenly released every week (although not exactly, but not too much), 37% of the time is a little more than two months. It has been a month since then, and the patient with procrastination, Mark, also said optimistically, "We still have time." But after being embarrassed by the wife, and after a "37% algorithm", he went home and looked at the room.
▲This is the house.
The next thing is this: We built a database, which has more than 30 houses that we have seen in more than two months, the address of the large-sized apartment, so small that I like the washing machine brand attached, no matter how small. Finished out. After more than a month, we entered the 37% stage and quickly started making decisions. After only three weeks, we got a favorite house for the auction. Thanks also to the "secretary algorithm"!
If the "secretarial algorithm" can be used to apply for employees and find a house, can it be used to solve some of the bigger problems in life, such as looking for an object?
The author of this book, Brian Christian, was very confused in his first love. When he was in college, he still maintained a long-distance relationship with a high school girl, but he could not avoid his own torture: What is the quality of this first love relationship? He has no ex-girlfriends to make comparisons, and there is not enough evidence to judge the quality of his first love. Brian confused to find a counselor to discuss. The counselor who is not used to telling him tells you that you must collect data first. Only if you have enough love experience and have seen all kinds of people can you decide whether this relationship is the best.
But how many relationships will it take to make sure you find the best person? After a thousand sails may not be, suddenly look back may also find that the best people have disappeared into the sea of people. After I finally learned how to love, some people will no longer miss it. Now that artificial intelligence can do everything, can algorithms be used to save humans struggling in painful love? is 37% a golden number in search of objects?
Michael Trick, a computer professor at Carnegie Mellon University, considered this issue when he was a graduate student. As a young mathematics doctor who is smart and hopes to find a girlfriend, he found that this love problem is actually a "secretary question." His calculation is like this: Assume that the age of 18 to 40 is the time to find the object, 37% of the time is 26.1, and he is at this age! He then found a girl who felt that he was beyond the past in all aspects and did not hesitate to propose to her. However, he was rejected.
After grief for love, Trick re-studied the secretarial algorithm. He found it much more complicated to find a girlfriend than to buy a house. The original secretarial algorithm had a huge assumption that your offer would be accepted by the other party. Even in our Bay area, home purchases are generally offered by high-priced people, and there is an intermediary who helps to value them. As long as they are determined to win, they can always make acceptable choices, and there is no big difference in the models. But girl, you never know what she wants, especially for steel straight male Ph.D. Programmers.
Trick then corrected the algorithm. He found that if the premise was corrected and the acceptance probability of the other party was changed to 50%, then you only had 25% of the time to collect the data, and the success rate was reduced from 37% to 25%. However, the most unfortunate thing is that he found himself too late, and 25% of the data collection time is already his history.
Let us see how the famous astronomer Kepler handled this object-seeking problem. In 1611, Kepler's wife died and he began to seriously look for remarriage. With the help of the enthusiastic matchmakers, he finally established 11 girls and started dating one by one. But when he dated to the fourth girl, he felt that he was looking for the right person. "I can actually stop looking." However, Kepler is a scum man who rides a horse to find a horse. He did not stop looking, but hung four girls, and continued to date with the remaining seven girls one by one. He finally found the fifth girl better. "I like her love, loyalty, family life, hard work, and her love for the stepchildren." He abandoned the other girls and proposed to the five girls, and the rest of the marriage was happy.
In the "secretary issue", Kepler surpassed the algorithmic restrictions, and with the reputation of the scum man, he adopted the strategy of "do not abandon and not give up." Under the premise that the girls are still willing to wait, reading thousands of sails to understand all the data and then make choices is of course the best strategy. But do you think mathematics men and doctors, can you let the girl wait for you? (Girls please polish your eyes and recognize the scum!)
But if you have to go straight ahead and give up when you choose, and wait until after a thousand red, and the object has a 50% chance of getting back together, when should you start making a choice? The omnipotent algorithm tells us that you can spend 61% of your time collecting data, and then choose the best choice in the top 61% and the best choice every time you come across, and your chances of getting the best partner-61%.
Note: this algorithm is also applicable to girls. Old girls don't give up looking for true love!)
Our mathematics doctor, Trick, finally got true love. After the marriage proposal was rejected, he continued to search until he found a German girl accepted the proposal eight years later. The book "Algorithms to live by" faithfully records his love story and recommends it to everyone who is single dog.
The above article was written in 2016. Now my baby peaches are already one and a half years old. My girlfriends have been discussing love and discussing buying a house. Now they are in the state of discussing kindergarten. When my female doctors and girlfriends complained about visiting the kindergarten every day and didn't know how to choose, I suddenly thought of the secretarial algorithm.
Bay Area Kindergarten is also extremely sought after. Basically, all mothers have to go through such a mental journey: in the face of many kindergartens with numerous characteristics, it is impossible to look at the kindergartens that they like, and they will find that the kindergartens are full in the morning, and the children are not ready to go to school at home. Mother's luck, go into a panic state, pray that as long as the child has learned, but then see the kindergartens that nobody cares about, it is impossible to get down, the mother loves to shed the house rather than resign at home. How can I guarantee to find a good kindergarten? Is it 37%, 25%, or 61%?
There is a fundamental difference in finding a kindergarten than looking for a house and looking for an object. Houses and objects are one-on-one choices. At a certain point in time, this object and house can only have one such customer. But kindergartens are different. Every kindergarten has dozens of hundreds of children. In the Internet age, the parents of these children kept evaluating kindergartens on the review network, social media, and WeChat groups. It may be difficult for you to find a boyfriend's ex-girlfriend to understand his love performance, but you can easily understand the quality of the kindergarten from the parents.
As explained in the book < Algorithms to live by >, this means that you have knowledge of the selection object in the secretarial algorithm. It's like you give all candidates an exam and get an average score for all candidates, and you only choose a secretary among candidates above the average score line. This obviously increases your success rate in getting the best secretary. The algorithm shows that your success rate can change directly from 37% to 58%: you have a 58% chance of getting the best secretary.
In fact this may also be used to find objects. Love can't be quantified, but money is always ok. If your goal is to find an object with income above the average salary, and do the due diligence work when inquiring about the other's income (which is probably easier than asking your ex-girlfriend), then the probability of getting the best object is 37%. Increased to 58%! True love is rare, money is easy, and the most painful thing in the world is this.
When I gave my daughter Ph.D.'s best friend science a 58% algorithm, she suddenly went through all the local kindergartens and compiled an extremely detailed database, almost more amazing than her doctoral thesis. Her baby finally entered a very good kindergarten and began a new life happily.
As for how do I find a kindergarten for peaches? Of course, it is a direct use of the database of female doctors and girlfriends! The book has its own gold house, the book has its own Yan Ruyu, and there are kindergartens in the book. I wish you all a happy life in using algorithms and living a happy life.
This article comes from WeChat official account: ID:PeachAtEden of Garden of Eden, author: fujiaa, Oxford University Doctor, Science Entrepreneur, Science Squirrel member, Science Popularization author. There is a little peach in the family.
* The article is the independent view of the author, does not represent the position of the tiger sniffing net. This article was published by the peach of the Garden of Eden, authorized by Tiger Sniff Network, and edited by Tiger Sniff Network. Reprint this article, please indicate the author's name in the text, keep the integrity of the article (including the tiger sniffing and other authorship information), and please attach the source (Hua Sniff Network) and this page link. Original link: https://www.huxiu.com/article/288827.html If the person is not reprinted according to the rules, the tiger sniffs the right to pursue the corresponding responsibility.
In the face of the future, you and I are still children, not to download the Tiger Sniff App and sniffing innovation!