Future Super Intelligent System

After 24 days of testing, Liu Fan showed the results to Zhao Jianguo and six other people he brought in the conference room. After half an hour of video playback, Liu Fan got up and stood in front of everyone.

"The demonstration just now should be relatively clear, and now I should say it is to make a comment." Liu Fan is now facing such an occasion, obviously a lot more comfortable, "The core factors of the problem of information leakage and transactions are two, one It is the inner ghost, and the second is the hacker. I can do very little about the inner ghost. Let me talk about how my big data processing system can effectively target the hacker.

I heard that several of you here today are the heads of the technical department, so everyone must know that, in fact, the hackers who have hacked the Pentagon like the movie are selected by thousands of miles. Over 90% of the hackers use similar methods And it's easy to learn. To put it plainly, do five things well, hide the IP, find the target vulnerability, implant the wood dragon to tamper with the data, get the target, and quickly flash people.

And when they do these five things, the most important thing is time. Under the authorization of Zhao Bureau, I studied the hacking detection system that the police are currently using. There is no problem with logic and execution. The problem lies in speed. This speed problem can also be divided into two parts, one is processing speed, and the other is immediacy.

So the most intuitive content I just showed is speed. Everyone should have seen the experimental comparison just now. The system I optimized can increase the data processing speed by more than three times. So how is the real-time hacking alarm mechanism shown in the second part of the video realized?

This thing is still related to the particularity of my underlying algorithm. The detection system we are using now uses trace information. It is said that more than 1,800 trace information have been found, right? But under my algorithmic calculations, I have found more than twenty kinds of Trojan horse implantation logic. With these implantation logics, the system can alarm when a hacker is under attack. "

"Sorry to interrupt." Just then, a man suddenly raised his hand.

"What's your problem?" Liu Fan is still very polite to the police. It's not important to be unconvinced. The key is respect for this profession.

"I'm really sorry." The man said sorry to everyone again, "The main problem has troubled me for a long time, and I just want to ask about this. Mr. Liu mentioned your algorithm, and I have done it before Research, many people now think that your algorithm can lead humans to the era of true artificial intelligence.

My personal understanding of real artificial intelligence, or your algorithm, is to rely on the powerful processing power of computers, so that computers have the same way of thinking as humans, and computers can break the limitations of the human brain. Similar to the existence of super human beings, to deal with problems that countless humans cannot handle. It also includes what you mentioned earlier about finding order in disorder. "

凡 Liu Fan thought for a while and nodded, "You can understand that too."

"Then I have a doubt here. If a computer can really be personified, does it mean that the computer can make mistakes like human beings at the same time that computers can think like humans? Throughout human history, humans often One of the mistakes is a blindfold. When human beings have a better understanding of the world, they will find that their previous understanding of the world is wrong.

So how do you make sure that the function or rule derived by your algorithm is correct? Although the current use of face recognition is indeed very effective, if the amount of data becomes larger and larger in the future, what should we do if these rules are wrong? Will there be some confusion at that time? "

"It seems that this police officer is like a man who thinks hard." Liu Fan smiled. "I can answer this. Your argument just now is that people may make mistakes like humans after computer anthropology. I also considered it. So there is a function introspection mechanism in my algorithm.

In simple terms, once the algorithm finds a function rule in the data, he needs to match the function rule to all related data once, and the correct 100% function rule will be retained. When the data increases, each additional data function The rules are matched once.

Take facial recognition, as more and more face information is entered, this function rule has been matching. Once this function rule does not match on a certain person, even if the accuracy rate is as high as 99.99%, The system will also immediately remove this function from the list of rules.

In other words, in the current big data environment, the rules in the rule list can be guaranteed to be 100% correct, and as long as the criminals we need to find are in this database, no accidents will occur. If an accident occurs, it is also in an extreme case, for example, a brand new face data has just entered this library, and the function does not match, and this person happens to be a criminal, which just inspired this wrong function, and is just being photographed Come down. "

"I see." The man nodded. "If this is the case, there is no need to worry about face recognition. With the development of technology in these years, we can cover almost everyone's face data.

But like this anti-hacking technology, the data is constantly expanding. Is there another possibility that the system shows that there are so-called rules, but because the rules are wrong, there are false positives? "

"False positives are indeed a possibility." Liu Fan did not shy away. "This was raised during the face recognition project. Although it is a very unlikely event, I also have enough plans for this problem.

If there is only a small probability of false positives, we can manually report the errors. If there are more than two false positives within a specific time, the reverse push algorithm stripping procedure I set will be triggered, self-checking and retaining the traditional technology.

But even if we only keep the traditional technology, we still have advantages in speed and accuracy. We have shown this before. "

"What if a large number of false positives occur at the same time?"

"Technology is just an auxiliary tool." It was here that Zhao Jianguo suddenly opened his mouth, his low voice with an inexplicable deterrent.

"Zhao Bureau, I just want to ask all the hidden dangers that may exist."

"I understand what you mean." Liu Fan hurriedly smiled and then relieved the atmosphere of sudden depression. "But I actually also take corresponding measures to address your concerns, such as the weight threshold response mechanism.

In all logics, there is actually a distribution of the proportion of the frequency of use. For example, the proportion of A logic is 80%, and the proportion of B logic may be 10%. So in the system alarm we received, where did it pass? A logic-triggered alarm will also have a weight distribution.

I mentioned before during the implementation of face recognition ~ www.readwn.com ~ dynamic features are still features. In too extreme cases, my technology will still be invalid. This time the hacking is also the same, although some are found. The Trojan implants logic, but if the Trojan implant method used by hackers just avoids this logic, then the advantage of this system in this regard will be lost.

In addition, crime itself is a small probability event, so we can make a time difference. We assume that the average alarm time is half an hour. If there are two false alarms in one hour, the system will automatically strip, so the one you just mentioned This kind of extreme phenomenon, first you have to have dozens or even hundreds of alarms in one hour, and at this time the system will display an abnormality. According to the weight ratio, priority is given to 10% of high-alarm alarms, and the remaining 90% will automatically start Stripping procedures, but not doing nothing, just shifting to traditional tracking procedures.

If there is no false positive upload after a certain time, subsequent system alarms will continue to be triggered according to the settings, which avoids wasting police resources and energy due to a large number of errors in the system in a short period of time.

And I have set dozens of plans like this in the system. To put it bluntly, your worry must be because these rules are derived by algorithms, so you think it may be uncontrollable. I have considered these issues and looked back. We can talk privately. "

Liu Fan said this, the previous man nodded with satisfaction, "Trouble Mr. Liu, I feel more relieved when you say so."

While the big brothers didn't notice, Fang Hua suddenly poked Wan Cuicui.

Wu Wancuicui was startled and looked at Fang Hua with a little anger, "What are you doing?"

Wu Fanghua pointed at Wan Cuicui's eyes and smiled, "You have stars in your eyes."

Wan Wancuicui: "..."