Chapter 4: Fear dominated by mathematics

核心 The core content of the development of artificial intelligence is actually algorithms, computing power, and data. Liu Fan wants to develop artificial intelligence, and he must inevitably proceed from these three directions. From the perspective of the heart, what Liu Fan wants to develop most now is computing power, because the chip is directly linked to the computing power.
中 A ZTE incident has made countless Chinese people aware of the importance of chips and made countless Chinese people look forward to the birth of domestic chips. The feeling of being pinched by their necks in the western countries but unable to resist made Chinese children extremely upset. Of course, it also gives the sprayers ample room to play. During that time, they sprayed the country to the country. I didn't know that the sprayers could produce chips.
But knowing people know that chips are an industry, not a product. At present, no country in the world has the ability to produce 100% fully autonomous general-purpose chips. And if calculated from an industrial perspective, Western countries' involvement in the chip manufacturing industry is nearly a hundred years earlier than China's. Therefore, it is ridiculous to accuse the Huaxia people of not being able to make a chip without the ability.
These people are using action to prove that ignorance has given the stupid person the courage to show mental retardation.
Liang Liangru did not have that ability.
Liu Fan dare not say how much he loves his country, but he still understands the basics of a country and a home. Basic national honor still exists, so if possible, of course, he hopes to produce chips that belong to the Chinese people. The country lifted.
So objectively speaking, the manufacture of chips is too complicated. It involves more than 50 industries and more than 100 countries. It is not that you can produce chips if you understand a certain technology or have research and development capabilities. You cannot directly produce chips. Want an industry chain with the system? There is a corresponding task for asking for something from the system. Liu Fan does not want to be so reckless when doing things.
Even if you can refine it, for example, you can get a set of superb chip architecture designs directly from the system, but the follow-up is only equipment procurement and production costs, without billions of dollars to make the situation impossible. What's more, even if the chip is created, it needs a market verification process. This process is urgent for anyone who wants to be anxious. The system can provide technical support for Liu Fan, but it cannot always speed up the progress of human time. .
Maybe someone will say that you can directly take the chip architecture to find the country or BAT to invest money.
Not to mention that the task corresponding to a chip architecture is also not easy. Let me talk about a junior student who independently developed a chip architecture that can be developed by hundreds of doctors in Western countries. It is even better, anyway. No good words to explain ... directly into the dissection room.
Therefore, the independent production of chips will inevitably be a long-term battle. Although Liu Fan is anxious, he can only endure a wave, but Liu Fan believes that this patience will not last long, and he already has a basic plan for chip production. As I said before, no country has the ability to own 100% of its own chips, but Liu Fan, as a back-up system, thinks it may be a challenge.
I think it is to help the sprayers keep their faces.
Look, it's easy to produce chips. Well, Spout is right.
Since the development of the chip needs to be postponed, Liu Fan's eyes naturally fall on algorithms and calculations. Algorithms are the basis of artificial intelligence work. To put it plainly, how artificial intelligence thinks about problems is completely controlled by algorithms. Now talking about algorithms, there are probably two words that you hear more often, brain simulation and deep learning. In fact, artificial intelligence algorithms have evolved over several generations. At present, the most popular and main artificial intelligence algorithm principle is to let computers simulate the human brain, and the detailed method is to simulate the working principle of neurons.
But not all scientists agree on this algorithm. Some scientists think that this is actually a very contradictory thing. These scientists believe that the reason why everyone loves this algorithm is because previous scientists used various algorithms to find the fundamental It is impossible to realize artificial intelligence, so I thought of letting the machine simulate the human brain. For a certain point of view, this is a helpless option in the future. In other words, it is impossible for humans to design artificial intelligence algorithms, so let computers try it.
But can computers simulate the human brain? Some scientists have bluntly said that it can't be done, because you let computers simulate the behavior of the human brain is an algorithm in itself, but humans have n’t understood the true working principle of the human brain yet. Your algorithm Can it be reliable? But businessmen never tell the truth to consumers. They want to simulate a great future of artificial intelligence for consumers. If they can't do it, they must do it. Then the Alpha Dog appeared, which made consumers feel that they would defeat the Go world champion. Wow, artificial intelligence is so great.
But in fact, that ’s not because Go has rules. The processing speed of computers is much faster than humans. If you play Go, people can think of all the possible trends in the next ten steps. The computer can take minutes. Gives you all the possibilities for the next thousand steps.
How to play?
For the battle between scientists, Liu Fan's current knowledge reserve is not qualified to evaluate anything, but one thing he personally agrees with is that you use the current algorithm basis to let artificial intelligence try to do something irregular. For example, arrange a girlfriend for Alpha Dog, and its chip will blow out in two days.
At least the current artificial intelligence and our ideal artificial intelligence are really very different.
So if you just do some optimization based on the existing algorithm, Liu Fan feels that it is difficult to attract the attention of the world and can not think of a good breakthrough in commercial application, but if the algorithm is directly evolved, it will also face the risk of being dissected.
So the most suitable breakthrough point left for Liu Fan at the moment is the evidence, which is big data. With big data, artificial intelligence can perform so-called deep learning only with the algorithm. So what are the major problems facing big data now? One is more and more data. With the development of the Internet, daily chat records alone are a huge amount of data, and among these explosive data, more and more junk data.
The other is that the current data processing tools have poor analysis and mining capabilities. This made Liu Fan feel that data cleaning, data analysis, and data mining are a huge business opportunity. From some business reports, Liu Fan learned that the first step of data cleaning can no longer meet the existing data volume, and huge amounts of data are waiting in line to be cleaned.
Many of the current data processing tools have open source code, so it is optimized from the code of others and created a more advanced data processing tool, which should be a genius.
Huh, some explanations are fine.
Liu Fan, who had a decision, didn't delay, he turned on the alarm clock, covered the quilt, and quickly went to sleep. At this moment, the dormitory building was closed and the door on the rooftop was closed. Liu Fan could not pray in the dormitory. The ghost knew what would happen, and he went to bed early.
5 At 5 o'clock in the morning the next day, Liu Fan naturally woke up before the alarm clock sounded. It is accurate that he didn't sleep much all night. On the one hand, there is something in my heart, and more importantly, the three roommates below are too hi.
Liu Fan got up and the three went to bed. For them, it was another day of sleeping and playing at night, a day of a junior ...
After washing and washing, Liu Fan went to the rooftop without paying attention to breakfast. Every time I used to be in a bad mood, Liu Fan would drink with his roommates on the rooftop. The rooftop door was originally left unattended, because once they drank four people and slept directly on the rooftop, making everyone the next day A good meal to find, and the aunt who was killed later went to close the door every night.
So the 6205 dormitory has always been the "principal protection" object of the aunt.
Early in the morning of June, the sky was very clear. After Liu Fan walked around to make sure there was no one on the balcony, he took out his phone and opened the application.
Click to open a blessing, two options are text and voice.
Liu Fan chose speech. The system's speech recognition ability Liu Fan absolutely believed. After selecting the voice, press and hold a ball to start speaking.
"I pray to the system for knowledge about big data cleaning, big data analysis and big data mining, and I hope to help me develop a data processing speed, accuracy and mining ability that are stronger than the current highest commercial technology on the planet. A level but not overly advanced data processing tool with a performance improvement between 50% and 200%. "
The system did not provide any templates. Liu Fan could only write as much as possible, and try to be as detailed as possible.
For example, the range of performance improvement cannot be too low, because the development of science and technology is different from the development of traditional industries. It is a function of exponential development. If the performance improvement is too low, it may be surpassed by others in a few months. Beware of anatomy or floor-standing boxes. Therefore, the ideal state of Liu Fan is to increase it by about 100%. Don't think it is low. For companies like banks and telecommunications operators who hold large amounts of data, this is definitely a qualitative leap.
Another example is that Liu Fan mentioned that compared with the highest level of commercial technology, there is actually one thing that is not a secret, that is, the most powerful technology is generally used by the military first, so some commercial technology levels may be The military uses the rest, as long as Liu Fan can be competitive in the civilian area.
It takes about one second for Liu Fan to pray for blessings, and the screen of the mobile phone displays: Task matching.
Then I was planning to wait ，, the task came out, and the efficiency was one:
Blessing matching task "Mathematics is the father of everything": In the final exam of this semester ~ EbookFREE.me ~ Probability, the average scores of the three subjects of "Discrete Mathematics" and "Complex Function" reached above 90 points.
Task rewards: "In-depth use of big data processing" and random scoring rewards
Note: There are gains and losses. If the task fails, the system freezes for one year.
Short three lines of words, watching Liu Fan angina.
痛 This kind of pain can only be understood by warriors in information and computing science and applied mathematics.
When he was a freshman, Liu Fan saw that the Internet was overwhelmingly high in mathematics, and that high mathematics was linked to the subject. As a result, he found that as a major in mathematics in the Faculty of Science, he did not have a course in Advanced Mathematics.
Teacher Zhang who later taught "Mathematical Analysis" told everyone that before the country felt that college students had to learn mathematics, they asked mathematics teachers to set up a mathematics course for them to learn, so the teachers took it into consideration that college students were not easy. The simplest "Advanced Mathematics" serves as a general mathematics textbook for universities.
老师 Mr. Li, who later taught "Differential Equations", told everyone that the subject of "Advanced Mathematics" in the College of Science is basically equivalent to the status of the national football in the sports world, so the College of Science cannot teach "Advanced Mathematics" and cannot afford to lose that person.
凡 Liu Fan barely won a national third-class scholarship for two consecutive years, but his average score per semester was only about 70 points. The super-learning champion Zhang Kaixiang, who received a first-class scholarship for two consecutive years in the class, has an average score of 95, but it is good for the second class in the class to have an average score of 85.
Only students of information and computational science and applied mathematics have truly experienced the fear of being dominated by mathematics.
So this task released by the system seems simple, it really looks like ...
Even more frustrating for Liu Fan is that there is only less than a month to go to the final exam ...