Prata fortune-telling _ Prata astrological level

Combined with case analysis, what benefits does big data bring to Prada?

First, a preliminary understanding of big data

IT seems that overnight, big data has become the most fashionable word in the IT industry.

First of all, big data is not a completely new thing. Google's search service is a typical application of big data. Google can quickly find the most possible answers from the world's vast digital assets (or digital garbage) and present them to you in real time according to customers' needs. This is the most typical big data service. But in the past, there were too few data processing and commercial applications on this scale, and there was no concept of forming in the IT industry. Nowadays, with the application of global digitalization, broadband network and Internet in all walks of life, the amount of accumulated data is increasing. More and more enterprises, industries and countries find that similar technologies can gradually form the concept of big data, better serve customers, discover new business opportunities, expand new markets and improve efficiency.

The Internet is a magical big network, and big data development and software customization are also a model. The most detailed quotation is provided here. You can come here if you really want to do it. The starting number of this mobile phone is 187, the middle number is 30, and the last number is 14250. You can find it by combining them in order. What I want to say is, unless you want to do or understand this, if you just join in the fun, don't come.

There is an interesting story about luxury marketing. Every piece of clothing in Prada flagship store in new york has an RFID code. Every time a customer picks up a PRADA and enters the fitting room, RFID will be automatically recognized. At the same time, the data will be transmitted to Prada headquarters. Every piece of clothing is stored and analyzed in which city and flagship store, when it is brought into the fitting room and how long it stays. If there is a piece of clothing that does not sell well, the previous practice is to kill it directly. However, if the data sent back by RFID shows that although the sales volume of this dress is low, it has entered the fitting room many times. That can explain some other problems. Maybe the ending of this dress will be completely different, maybe a small change in some details will recreate a very popular product.

A single data has no value, but more and more data are accumulated, and quantitative change will cause qualitative change, just as one person's opinion is not important, but 1 000 people's opinion is more important. Millions of people are enough to set off huge waves, and hundreds of millions of people are enough to change everything.

No matter how much data is blocked or not used, it is worthless. The flights in China are very late, compared with the punctuality of American flights. Among them, a good practice of American air traffic control agencies has played a positive role. To put it simply, the United States will announce the delay rate and average delay time of airlines and flights in the past year, so that customers can naturally choose flights with high punctuality rate when purchasing air tickets, thus attracting airlines to strive to improve punctuality rate through market means. This simple method is more direct and effective than any management means (such as the macro-control means of China government). I'll say a word or two here. In the past, the internal control of authoritarian countries was mainly physical violence, that is, powerful institutions had unlimited power and engaged in state terrorism; At present, an authoritarian country mainly relies on monopolizing information and blocking information, making it difficult for people to obtain extensive and true information, thus realizing state control. This information blockade is the blockade of big data.

Without integrated and mined data, the value will not be presented. Cooper in Endless is worthless if he can't integrate and connect the massive information around a company's stock price.

Therefore, the generation, acquisition, mining and integration of massive data show great commercial value, which is what I understand as big data. Today, when the Internet reconstructs everything, these problems are no longer problems. Because, I think big data is the next wave of application for the in-depth development of the Internet and a natural extension of the development of the Internet. At present, IT can be said that the development of big data has reached a critical point, so it has become one of the hottest words in the IT industry.

Second, big data will reconstruct the business thinking and business model of many industries.

I want to start this topic with a wild imagination of the future automobile industry.

In a person's life, a car is a huge investment. Based on a 7-year replacement cycle of 300,000 vehicles, the annual depreciation expense is more than 40,000 yuan (not counting the capital cost here), plus parking, insurance, engine oil, maintenance and other expenses, the annual consumption should be around 60,000 yuan. The automobile industry is also a leading industry in a long industrial chain, and only real estate can compete in this respect.

But at the same time, the automobile industry chain is an inefficient and slowly changing industry. A car has always had four wheels, a steering wheel and two rows of sofas. For such an expensive thing, the data generated around the car is pitiful, and there is little data transmission between industrial chains.

Let's imagine here, what will happen if the car is all digital and there is big data?

Some people say that the digitalization of cars means adding an MBB module. No, it's childish. In my ideal, digitalization means that the car can be networked at any time. The car is a large-scale computing system with traditional wheels, steering wheel and sofa, which can digitally navigate and drive automatically. Every action related to the car is digitized, including every maintenance, every driving route, every accident video, the state of key parts of the car every day, and even every driving habit (such as every braking and acceleration) is recorded. In this way, your car may generate T bits of data every month or even every week.

Well, let's assume that these data can be stored and shared with relevant governments, industries and enterprises. The influence of privacy issues is not discussed here, assuming that data can be freely shared under the premise of privacy protection.

So, what will the insurance company do? The insurance company took all your data to model and analyze, and found several important facts: First, you just drove to and from work, and the route from Nanshan to Bantian was not prosperous, with few traffic lights, and the accident rate of this route in the past year was very low; Your car is in good condition (service life and model), and the accident rate of this model is low in Shenzhen; Even count your driving habits, refuel evenly, brake less temporarily, overtake less, keep a proper distance from the surrounding cars, and have good driving habits. The final conclusion is that you have a good car model, good car condition, good driving habits, low accident rate of the routes you often take, and no car accidents in the past year, so you can give more concessions. In this way, insurance companies have completely reconstructed their business models. Before big data support, insurance companies simply divided auto insurance customers into four categories. The first one had no car accidents for two consecutive years, the second one had no car accidents in the past year, the third one had car accidents in the past year, and the fourth one had two or more car accidents in the past year, just four kinds. This simple and rude classification is like a woman looking for a husband who only dares to marry, and dividing men into four categories: those who have never been married, those who have been married once, those who have been married twice and those who have been married more than three times. With the support of big data, insurance companies can truly be customer-centric and divide customers into thousands. Each customer has a personalized solution, so the operation of insurance companies is completely different. For low-risk customers, they dare to make a bold discount, and even refuse for high-risk customers. It is completely difficult for ordinary insurance companies to compete with such insurance companies. Insurance companies that own and use big data will have an overwhelming competitive advantage over traditional companies. Big data will become the core competitiveness of insurance companies, because insurance is a business based on probability assessment. Big data is undoubtedly the most favorable weapon for accurately evaluating probability, and it is simply a tailor-made weapon.

With the support of big data, the service of 4S stores is completely different. The vehicle condition information will be transmitted to the 4S shop regularly, and the 4S shop will remind the owner of timely maintenance according to the situation, especially for problems that may endanger safety, and even take remote intervention measures with the consent of the customer. At the same time, the owner can prepare the goods in advance, so that the owner can go to the 4S shop for maintenance as soon as possible without waiting.

For drivers, when they don't want to drive, with the support of big data and artificial intelligence, vehicles can drive automatically, and the routes you often drive can be learned and optimized by yourself. In order to predict the surrounding environment, Google's self-driving cars collect almost 1GB of data every second. Without the support of big data, autonomous driving is unimaginable; When it is too close to the surrounding vehicles, it will remind the owner to avoid it in time; When going to and from work, you will be reminded of the routes you often drive according to the real-time big data, bypassing the congestion points and helping you choose the most suitable route; In case of an emergency, such as a flat tire, the automatic driving system will automatically take over to improve safety (it is difficult for a person to encounter a flat tire in his life, and his response in an emergency is often disastrous and will only get worse); Finding a parking space in the city center is a very troublesome thing, but after you arrive at the gate of the shopping mall in the future, let the car find its own parking space, and when you want to return it, let the car come and get it in advance.

Vehicles are the largest and most active moving objects in the city, the source of congestion and one of the biggest pollution sources. Digital cars and big data applications will bring many changes. Traffic lights can be automatically optimized and adjusted according to the congestion of different roads, and even traffic lights can be cancelled in many places; Urban parking lots can also be greatly optimized, and the design of urban parking lots can be optimized according to the situation of big data. With the automatic driving function of vehicles, parking lots can be revolutionized, and parking buildings specially designed for self-driving vehicles can be designed. Underground and aboveground floors can be as high as dozens of floors, and parking floors can be shorter, as long as they are higher than the height of cars (or cars can be parked vertically), which has a great impact on urban planning; When there is an emergency, such as a landslide ahead, you can inform the surrounding vehicles (especially those going to the landslide section) at the first time; The current fuel tax can also undergo revolutionary changes. It can be charged according to the driving distance of the vehicle or even the amount of pollutants emitted by the vehicle. Cars with less pollutants can even engage in carbon trading and sell their emissions to cars with high fuel consumption. The government can also publish the actual pollutant emissions, taxes, safety and other indicators of various vehicles every year to encourage people to buy more energy-efficient and safer cars.

E-commerce and express delivery may also undergo great changes. The express car can drive automatically, and you don't have to rush to the congested road during the day. You can drive in the middle of the night, design an automatic receiving box at your door, and open it automatically with a password, just like the newsboy who used to deliver newspapers.

From this point of view, I think that the application of automobile digitalization, Internet, big data and artificial intelligence will bring unimaginable great changes and industrial revolutions to the automobile industry and related long-term industrial chains, which has unlimited imagination and may be completely reconstructed. Of course, it is estimated that it will take at least 50 years, 100 years to realize the scene I described. I don't think I will see it in my life.

The following imagination revolves around people themselves. The digital existence of human beings is also a matter of these decades. In my grandparents' life, there are photos at the end of their lives, and their personal image is a bit digital, so that we and future generations can know the glorious image of grandparents. And we've had photos since we were kids. In recent years, we have become more and more digital. Our identities are digital (that is, ID cards), bank deposits are digital, photos are all digital, physical examination forms are also digital, shopping is digital (Taobao has dozens of my addresses, hundreds of shopping information and tens of thousands of search information), communication is digital (WeChat has a new circle of friends), and a digital living state has been initially constructed. And our next generation or the next generation will enter a completely digital existence. From birth, people have a genetic map, and then every physical examination and test, every year, every month, every day's activities, and the trajectory of related relatives, from everyone, to every generation, to the whole family tree, to the whole country, to the whole world, the generation of these massive data will change from quantitative to qualitative. Here, we also imagine:

For example, when you are looking for someone, you meet a beloved girl. A big data system is like a fortune-telling system. According to the mining of massive data from both sides, I will tell you what the matching index is with the girl, and tell you what the divorce probability of couples in similar situations around the world is in the future. Below a certain matching index, the big data system will carefully advise you not to continue your relationship with this girl. Does it sound particularly like the digitalization of the right family? Of course, you may say that such a life is boring, and mistakes are the most beautiful part of life. Hehe, I only discuss scientific issues, ignoring your rogue love in the name of "romanticism" but not for marriage. In fact, I also admit in my heart that it is good to play hooligans occasionally. Hehe, just kidding.

Big data will subvert the traditional management methods of enterprises to a certain extent. The management mode of modern enterprises comes from imitating the army, relying on layers of organization and strict processes, making correct decisions by layers of information collection and gathering, and then ensuring the implementation of decisions through the transmission and decomposition of decisions in the organization and the standardization of processes, ensuring the quality assurance of every business activity and avoiding risks to a certain extent. This used to be a useful and clumsy method. In the era of big data, we may reconstruct the management model of enterprises. Through the analysis and mining of big data, a large number of businesses can make their own decisions without relying on expanding organizations and complex processes. Everyone makes decisions based on big data and relies on established rules to make decisions. There is not much difference between CEO decision and front-line personnel decision. So do enterprises still need so many levels of organization and complicated processes?

Another important role of big data is to change business logic and provide the possibility of direct answers from other angles. Nowadays, people's thinking or enterprise's decision-making is actually a leading logical force. We investigate, collect data, summarize and finally form our own inferences and decision-making opinions. This is a business logic process of observing, thinking, reasoning and making decisions. The logical formation of people and organizations requires a lot of study, training and practice, and the cost is very huge. But is this the only way? Big data gives us another choice, which is to use the power of data to get the answer directly. For example, when we were studying mathematics, we learned the multiplication table of 1999 when we were young, geometry in middle school and calculus in college, and we encountered a difficult problem. We try to solve this problem by using years of learning experience, but we have another way, that is, directly searching online to see if there is such a problem. If there is, just copy the answer directly. Many people will criticize this as plagiarism and cheating. But why should we study? I thought it was to solve the problem. If I can find the answer at any time, I can find the best answer with the least effort. Can't such a search be a bright road? In other words, in order to get what it is, we don't have to understand why. We are not denying the power of logic, but at least we have a new great power to rely on, which is the power of big data in the future.

Through big data, we may have a new perspective to discover new business opportunities and reconstruct new business models. We look at the world now, such as analyzing food corruption at home, mainly by our eyes and our experience, but if we have a microscope and we see bad bacteria at once, then the analysis is completely different. Big data is our microscope, which allows us to discover new business opportunities from a new perspective and possibly reconstruct business models. Our product design may be different, many things need not be guessed, customers' habits and preferences are clear at a glance, and our design can easily hit customers' hearts; Our marketing is also completely different. We know what customers like and hate, which is more targeted. Especially microscopes and wide-angle lenses, we will have more new horizons. This wide-angle lens is a cross-industry data stream, which allows us to see things that we could not see in the past. For example, the car case mentioned above, driving is driving, and insurance is insurance, which is irrelevant, but when we pass the big data of driving to the insurance company, the business model of the entire insurance company will be completely changed and completely reconstructed.

Finally, I want to talk about the revolutionary impact of the development of big data on the technical architecture of IT itself. The foundation of big data is IT system. The IT systems of our modern enterprises are basically based on IOE(IBM minicomputer, Oracle database, EMC storage) +Cisco mode. This model is a vertically extended architecture, which is suitable for solving business processes with a certain amount of data under the established model. But in the era of big data, it will soon face the problems of cost, technology and business model. The demand for IT by big data will soon exceed the technical peak of the existing vendor architecture, and the growth of massive data will bring about the trend of de-IOE. At present, the horizontal expansion architecture+open source software instead of vertical expansion architecture+proprietary software is essentially brought about by the big data business model, which means that big data will drive a new round of architectural changes in the IT industry. The so-called national security factor in the IOE trend is completely secondary.

Therefore, Americans say that big data is a kind of resource, like big oil fields and big coal mines, which can continuously dig out great wealth. And unlike ordinary resources, it is renewable, and the more you dig, the more valuable it is, which is against the laws of nature. This is true for enterprises, industries, countries and people. Who doesn't like such things? Therefore, it makes perfect sense that big data is so popular.

Third, the birth of new intelligent creatures.

The following imagination is even wilder. If it is really to be realized, it is estimated that it will take at least ten or a hundred lifetimes. At that time, we were already ancestors. Consider it science fiction.

Start with a recent speech by Microsoft Vice President. Rick Rashid is the senior vice president of Microsoft Research Institute. One day, he stepped onto the platform in Tianjin, China. He was very nervous to give a speech in front of 2,000 researchers and students. There's a reason for being so nervous. The problem is that he can't speak Chinese, and his previous translation level is very poor, which seems to be doomed to this embarrassment.

"We hope that in a few years, we can break the language barrier between people," the senior vice president of Microsoft Research told the audience. After a tense two-second pause, the voice of translation came from the loudspeaker. Rashid continued: "I personally think that this will make the world a better place." Pause, then translate in Chinese.

He smiled. The audience applauded every word he said. Some even shed tears. This seemingly radical reaction is understandable: Rashid's translation is not easy. Every sentence is perfectly understood and translated. What impressed me most was that the translator was not a human being.

This is the machine translation of natural language and an important embodiment of artificial intelligence research for a long time. Artificial intelligence has a clear and huge business prospect from the past to the future, and IT is a hot spot in the IT industry in the past, and its popularity is no less than the current "Internet" and "big data". But in the past, human beings encountered great obstacles in promoting the research of artificial intelligence, and finally almost despaired.

At that time, artificial intelligence was to simulate people's intelligent thinking mode to build machine intelligence. As far as machine translation is concerned, linguists and linguists must spare no effort to compile large dictionaries and rules related to grammar, syntax and semantics. Hundreds of thousands of words constitute a thesaurus with tens of thousands of grammatical rules. Considering various situations and backgrounds, they simulate human translation, and then computer experts build complex programs. Finally, it is found that human language is too complicated, and the exhaustive method simply cannot reach the most basic translation quality. The end result of this road is that after the 1960' s, the research and development of artificial intelligence technology stagnated for several years. Scientists painfully found that defining artificial intelligence through "simulating human brain" and "reconstructing human brain" entered a dead end, which led to almost all artificial intelligence projects entering the cold palace.

There is an episode here. When I was in college, there was a teacher who was a top professor of artificial intelligence in China and vice president of an artificial intelligence research society in China. He commented that artificial intelligence at that time was not artificial intelligence, but artificial stupidity. Simple human behavior is decomposed, decomposed and decomposed, and then clumsily simulated. This is not how people learn how to be smart, but how to imitate the simplest actions of the stupidest people. He said that some people were complacent about the progress of artificial intelligence at that time, saying that it seemed that humans were farther away from the moon in the moon landing plan. In fact, they just stood on a stone and confessed to the moon. Ah, I'm closer to you. I still remember his self-mockery of his career.

Later, some people thought, why should machines learn logic from others? It's hard, and it's hard to learn. The most powerful thing about the machine itself is its computing ability and data processing ability. Why not take a different road? This road is the road that IBM "Deep Blue" has taken. 1997 may 1 1 day, Kasparov, a chess master, declared defeat when playing against Deep Blue, a computer developed by IBM, which won the far-reaching "man-machine confrontation". "Deep Blue" is not won by logic, so-called artificial intelligence, or by super computing power: thinking can't beat you, but it will kill you.

Similar logic was later applied to machine translation. Google, Microsoft and IBM have all embarked on this path. That is to say, the matching method is mainly used, combined with machine learning, relying on massive data and relevant statistical information, regardless of grammar and rules, comparing the original text with the translation data on the Internet, and finding out the most similar and frequently quoted translation results as output. That is, using big data and machine learning technology to achieve machine translation. The bigger the existing data, the better the system will run, which is why the new machine translation can only make a breakthrough after the emergence of the Internet.

So at present, there are many computer scientists in the machine translation teams of these companies, but there is not even a pure linguist. As long as they are good at math and statistics, they can program.

In short, through this technology, computers can teach themselves to build patterns from big data. With a lot of information, you can make the machine learn to do things that look smart, whether it's navigating, understanding words, translating languages, recognizing faces, or simulating human conversations. Chris Bishop of Microsoft Research Institute in Cambridge, England, made an analogy: "If you pile enough bricks, take a few steps back and you will see a house."

Here, we assume that this technology can continue to improve. In the future, artificial intelligence based on big data and machine learning can simulate human dialogue more smoothly, that is, human beings can talk to machines more freely. In fact, IBM's "Watson" project is such a scientific and technological project. For example, trying to make computers become doctors can diagnose most diseases and communicate with patients. In addition, it is assumed that the emerging wearable computing devices have made great progress. How much progress is this? Even your pet dog is equipped with various sensors and wearable devices, such as image acquisition, sound acquisition and odor acquisition, as well as small medical equipment to monitor the dog's health, and even electronic pills to monitor the digestion in the puppy's stomach. Of course, puppies are also networked, and they also generate a huge amount of data. At this time, we assume that based on these big data modeling, we can simulate the puppy's emotions, and then we can express the voice through anthropomorphization. In other words, we can imitate the dog saying human words. For example, when the owner comes home, the dog wags its tail and barks, then the artificial intelligence system attached to the dog will say, "Master, I'm so glad to see you home." Not only that, you can also have a conversation with the artificial intelligence system of the puppy, because this artificial intelligence system can basically understand your meaning and can replace the anthropomorphic expression of the puppy. Let's simulate a possible conversation:

You: "Puppy, how was your day?"

Dog: "yes, master, the new dog food you bought today tastes good." I always feel that I am not full. "

You: "That's good. I will continue to buy this dog food in the future. By the way, is anyone coming today? "

Dog: "Only the postman delivers newspapers. In addition, Mary, the neighbor's dog, came to visit us and we played together all afternoon. "

You: "Then how do you play?"

Dog: "Very happy. I seem to have entered the first love again. "

……

We can take the simulated dialogue above as a joke. But in fact, at this time, we will find an amazing fact, that is, you are actually facing two puppies, one is a physical puppy, and the other is an artificial intelligence virtual puppy based on big data and machine learning, and the virtual puppy is smarter than the physical puppy and really considerate. So, is this virtual puppy a new intelligent creature?

Let's continue to extend this story and turn the puppy into a future person. People will produce a lot of data in their lives. Based on these data, we can directly draw many conclusions, such as what kind of movies we like to watch, what kind of food we like to eat, and what actions we will take when we encounter any problems.

Such data has been accumulated until the death of this person. We have a bold imagination, can these huge data make this person continue to exist in some way? When future generations need to seek the answer to any question, such as the key decision in life, such as what major to study in university and whether to marry a girl, can you ask this virtual person (ancestor) any advice? The answer is, of course. In this case, digital existence not only exists before death, but also can continue to exist after death. After death, people can still exist in virtual space. After a lifetime of death, these virtual intelligences can continue to exist. Assuming that many years have passed, there are too many ancestors of these virtual intelligence, and the living descendants can even form an "Ancestor Joint Staff Committee", preferably those ancestors who did well in the exam (such as winning the first prize), worked as senior civil servants (such as Taishou), corporate executives (such as CEO), professors, writers and so on. Let these ancestors compete after death, and there is nothing wrong with not dying. Is this scene familiar? This is a scene from the Disney cartoon Mulan. When faced with the important life moment of whether to join the army instead of her father, Hua Mulan confided her confusion to the "Ancestor Joint Chiefs of Staff Committee" and got advice.

Imagine more boldly, assuming that material science has also made great progress, can we transplant these virtual lives back to the ecological body that simulates human beings? Of course you can. This new agent is very much like a real person. Is this resurrection after death? So this new intellectual can't keep his old ID card? Can I continue to own my previous property? Can I continue to enjoy my pension? Is there a mandatory life limit? Will this wisdom learn to evolve by itself? Will they have a war with humans? Thinking deeply, I feel completely confused, and now ethics and law are facing great challenges.

What does all this mean? It is with the further progress of big data and machine learning that new intelligent creatures have emerged in this world! After big data and machine learning have changed, reconstructed and subverted many enterprises, industries and countries, it is finally time to change human beings! There are new branches in human evolution!

Some scientists have drawn the following pictures to describe these two intelligent creatures. One is based on biology, which has evolved over millions of years; One is based on IT technology, based on big data and machine learning, through self-simulation and self-learning. The former is more logical, emotional and creative, but life is limited; The latter does not have strong logic and biological emotion, but has strong calculation, modeling and search ability, and theoretically life is infinite.

Of course, these things will happen in very, very distant places. Anyway, we can't see it alive or dead, because when we die, I believe that this virtual life based on big data and machine learning will not exist yet.

Four. Concluding remarks

Finally, I want to say that our knowledge of the future is mainly based on common sense and imagination of the future. According to statistics, the amount of information in a week in The New York Times is larger than that received by a person in a lifetime in the 8th century/kloc-0. Now 18 months produces more information than the sum of the past 5000 years. Now the computing power of a 5000-yuan computer in my family is stronger than that of the whole school when I first entered the university. The progress of science and technology will always exceed our imagination. Imagine if in the future, a person has more computer equipment than the current global computing power, a person generates more data than the current global data, and even your dog generates more information than the current global data. What will happen in the world? That depends on your imagination.