News classification
Contact us
- Add: No. 9, North Fourth Ring Road, Haidian District, Beijing. It mainly includes face recognition, living detection, ID card recognition, bank card recognition, business card recognition, license plate recognition, OCR recognition, and intelligent recognition technology.
- Tel: 13146317170 廖经理
- Fax:
- Email: 398017534@qq.com
The application of artificial intelligence in the field of network security
The application of artificial intelligence in the field of network security
What is intelligence?
Before we discuss the topic of artificial intelligence, let's first define what is intelligence. Intelligence is quite complex in the broad sense, and there are fierce arguments in many aspects of science and philosophy. But in this article, I provide the following definition.
I have two important ideas about intelligence. First, many scientists believe that human intelligence is rooted in how the brain discovers and stores relevant stratification in a variety of different types of sensory data. For example, when you see a "Gaurav-iPhone" in a network title in a captured packet or log file, you will naturally think that it is likely to be your friend Gaurav's iPhone. You will unknowfully connect the knowledge of the names of the colleagues with the knowledge of the common types of equipment. In life, you will update these two models from time to time, and will be influenced by multi - source multimedia sensory inputs that include Apple ads, TV programs, email, articles, and hallways. You can try to compare the process with the traditional unscrupulously unscrupulously string matching program, and stick to the differences in the flexibility of input and the accuracy of the output.
Second, intelligence is a prediction, which is a way to deal with problems. For example, your eyes are trying to see everything it can see, and at the same time, the brain sends prediction information to the eye according to what it is expected to see in the eye. This prediction mechanism "fills" what is not recognized, that is why you cannot recognize the cause of visual blindness. This mechanism also allows you to walk in the dark bedroom in the evening without tripping: your brain sends signals to the motor nervous system and provides a model for the muscles to walk.
Conventional artificial intelligence and narrow sense artificial intelligence
The concept of artificial intelligence was first proposed by computer scientists who went beyond traditional procedures in 1950s. They have been inspired by super intelligent programs that are similar to humans, such as R2D2 and C-3PO in "Star Wars", and supercomputers in Superman III, which are conventional artificial intelligence. Conventional artificial intelligence does not exist today. We do not know how to imitate human brain to stop work, so that we do not know a small part of intelligence that imitates it.
The artificial intelligence existing today can be called narrow artificial intelligence. In the past, many useful products were used in narrow sense artificial intelligence. They could perform some tasks in quality and quantity, so that they could do better than human beings. For example, Amazon's Alexa, which has a limited range of input, does not leave a variety of narrow artificial intelligence technologies to accomplish some tasks, which makes it wrong to think that it is intelligent. The world champion of chess and go is also the application of narrow sense artificial intelligence. These narrow artificial intelligence systems have three intelligent elements discussed earlier: the knowledge of specific categories, the mechanism for obtaining new knowledge, and the mechanism to use them.
At present, there are several ways to deal with the problems of network security through narrow AI. Of course, the security robot that can pass the Turing test and replace the security team members does not exist, but the tools based on the narrow artificial intelligence can detect the threats and vulnerabilities early, and can better balance the security situation than most people.
The difference between AI, machine learning, expert system and deep learning
Machine learning is the application of resolution algorithm and the first step in the acquisition process of knowledge. It was produced in the process of artificial intelligence in 1960s. Machine learning can be regarded as an algorithm focusing on learning. Instead of writing specific computer instructions to complete a task, a computer uses a lot of data to stop "exercise", so that it can learn how to perform a task. Samples for exercise can be provided either externally or at the previous stage of the knowledge discovery process.
It
Over the years, many machine learning algorithms have been presented, including decision tree, resolution logic, clustering, Bayesian network and artificial neural network. AI is closely related to statistics, so that they stack each other.
Machine learning is thought to be born out of an expert system, but unlike it, the expert system deals with problems based on fuzzy rules based reasoning based on the well prepared knowledge system (rules). The expert system was touted as the most successful case of artificial intelligence in the 1980s. The principle behind the expert system is that the intelligent system gets the talent from the knowledge they have, rather than getting the talent from the specific reasoning strategy they use. In short, the expert system is learned, but not self-contained. They need human programmers or operators to make them smarter. But suppose we stop judging from the definition of intelligence, they are not smart.
Go back to the system that will learn. Machine learning is difficult, because the way of association between data in multiple dimensions is a difficult problem. This is a big data and computing intensive problem. From time to time, the human brain obtains a large number of sensory data from a large number of sources and overpasses many dimensions, slowly improving its model, and then arriving at the intelligence and professional knowledge of the skilled staff of the network security team.
Before we discuss the topic of artificial intelligence, let's first define what is intelligence. Intelligence is quite complex in the broad sense, and there are fierce arguments in many aspects of science and philosophy. But in this article, I provide the following definition.
I have two important ideas about intelligence. First, many scientists believe that human intelligence is rooted in how the brain discovers and stores relevant stratification in a variety of different types of sensory data. For example, when you see a "Gaurav-iPhone" in a network title in a captured packet or log file, you will naturally think that it is likely to be your friend Gaurav's iPhone. You will unknowfully connect the knowledge of the names of the colleagues with the knowledge of the common types of equipment. In life, you will update these two models from time to time, and will be influenced by multi - source multimedia sensory inputs that include Apple ads, TV programs, email, articles, and hallways. You can try to compare the process with the traditional unscrupulously unscrupulously string matching program, and stick to the differences in the flexibility of input and the accuracy of the output.
Second, intelligence is a prediction, which is a way to deal with problems. For example, your eyes are trying to see everything it can see, and at the same time, the brain sends prediction information to the eye according to what it is expected to see in the eye. This prediction mechanism "fills" what is not recognized, that is why you cannot recognize the cause of visual blindness. This mechanism also allows you to walk in the dark bedroom in the evening without tripping: your brain sends signals to the motor nervous system and provides a model for the muscles to walk.
Conventional artificial intelligence and narrow sense artificial intelligence
The concept of artificial intelligence was first proposed by computer scientists who went beyond traditional procedures in 1950s. They have been inspired by super intelligent programs that are similar to humans, such as R2D2 and C-3PO in "Star Wars", and supercomputers in Superman III, which are conventional artificial intelligence. Conventional artificial intelligence does not exist today. We do not know how to imitate human brain to stop work, so that we do not know a small part of intelligence that imitates it.
The artificial intelligence existing today can be called narrow artificial intelligence. In the past, many useful products were used in narrow sense artificial intelligence. They could perform some tasks in quality and quantity, so that they could do better than human beings. For example, Amazon's Alexa, which has a limited range of input, does not leave a variety of narrow artificial intelligence technologies to accomplish some tasks, which makes it wrong to think that it is intelligent. The world champion of chess and go is also the application of narrow sense artificial intelligence. These narrow artificial intelligence systems have three intelligent elements discussed earlier: the knowledge of specific categories, the mechanism for obtaining new knowledge, and the mechanism to use them.
At present, there are several ways to deal with the problems of network security through narrow AI. Of course, the security robot that can pass the Turing test and replace the security team members does not exist, but the tools based on the narrow artificial intelligence can detect the threats and vulnerabilities early, and can better balance the security situation than most people.
The difference between AI, machine learning, expert system and deep learning
Machine learning is the application of resolution algorithm and the first step in the acquisition process of knowledge. It was produced in the process of artificial intelligence in 1960s. Machine learning can be regarded as an algorithm focusing on learning. Instead of writing specific computer instructions to complete a task, a computer uses a lot of data to stop "exercise", so that it can learn how to perform a task. Samples for exercise can be provided either externally or at the previous stage of the knowledge discovery process.
It
Over the years, many machine learning algorithms have been presented, including decision tree, resolution logic, clustering, Bayesian network and artificial neural network. AI is closely related to statistics, so that they stack each other.
Machine learning is thought to be born out of an expert system, but unlike it, the expert system deals with problems based on fuzzy rules based reasoning based on the well prepared knowledge system (rules). The expert system was touted as the most successful case of artificial intelligence in the 1980s. The principle behind the expert system is that the intelligent system gets the talent from the knowledge they have, rather than getting the talent from the specific reasoning strategy they use. In short, the expert system is learned, but not self-contained. They need human programmers or operators to make them smarter. But suppose we stop judging from the definition of intelligence, they are not smart.
Go back to the system that will learn. Machine learning is difficult, because the way of association between data in multiple dimensions is a difficult problem. This is a big data and computing intensive problem. From time to time, the human brain obtains a large number of sensory data from a large number of sources and overpasses many dimensions, slowly improving its model, and then arriving at the intelligence and professional knowledge of the skilled staff of the network security team.