How Do I Become a Computer Support Specialist?
The computer expert system is the only software system with commercial value in the field of artificial intelligence. Its basic principle is that it can perform symbolic calculus based on predicate logic, that is, it recognizes symbols and the order in which they appear, searches for matches in the library and makes corresponding substitutions, and simple logical operations.
Computer expert system
Right!
- Chinese name
- Computer expert system
- Foreign name
- Computer expert system
- Brief introduction
- A set of program software
- Application area
- Artificial Intelligence
- The computer expert system is the only software system with commercial value in the field of artificial intelligence. Its basic principle is that it can perform symbolic calculus based on predicate logic, that is, it recognizes symbols and the order in which they appear, searches for matches in the library and makes corresponding substitutions, and simple logical operations.
- Obviously, an expert system must have a knowledge base to "remember" the knowledge of human experts. There is also an inference engine, which can draw conclusions based on the knowledge (facts and rules) in the knowledge base based on the input that requires a solution. Machines do not know any numbers and knowledge. How can it be deduced?
- For example, "Zhang San likes chatting" is a fact, it can be stored in the machine's memory first, the format is
- Like (Zhang San, chat)
- "Like" is called a predicate, which is a string of symbols. Machines can record many such facts, such as:
- Like (Li Si, swimming)
- Like (Li Si, chat)
- Like (Wang Wu, football) ...
- In addition to facts, the machine can remember rules. For example: Zhao Liu's friends must have the same hobbies as him, like chatting and swimming. Can be expressed as:
- Friends (Zhao Liu, X): One likes (X, chat), likes (X, swimming)
- Now let the machine infer, who is Zhao Liu's friend, namely:
- ? A friend (Zhao Liu, who)
- The machine first finds the rule head with the predicates "friends", and obtains "who" X should be. Then, according to the predicate at the end of the rule, it checks what facts are stored in the database and finds out whether it is "chat" or "swim" The match was Li Si, and it was determined that "who" should be "Li Si".
- From this simple example, it can be seen that as long as we arrange the knowledge in a formal symbol string, the machine can infer by matching the symbol string. In fact, the more rules, the more "smart" a machine is.
- Therefore, to establish an expert system, you need to exchange opinions with domain experts, and repeatedly formalize the expert knowledge and put it into the database. Obviously, the better the expression of knowledge, the easier it is to reason, and the more correct the conclusion is. There are many methods such as production rules, first-order predicates, frameworks, semantic networks, and scripts. The reasoning mechanisms are different. There is not yet a unified representation and reasoning machine.
- Expert systems that are not inferior to human experts have appeared in the fields of medicine, detection and control, aerospace, economics, finance, molecular chemistry, and management. For example, the famous Chinese medical diagnosis and treatment expert system GUDES-1 can raise tens of thousands of clinics in China to the level of famous doctors. The social and economic benefits of the expert system can be imagined.
- In the 1990s, commercial commodity expert system development tools appeared, such as CLIPS. Use this software tool to quickly develop a dedicated expert system.