Detailed · Deep study
1What is AI?
A branch of computer science dealing with the simulation of intelligent behaviours by computers.
| Characteristic | Meaning |
|---|---|
| data and rules | AI has a collection of data and the rules for using that data |
| reasoning | it can use the data and rules to reach conclusions, make decisions or solve problems |
| learning and adapting | it can include the ability to learn from new data or experience and change its behaviour |
Examples: voice assistants, chatbots, face recognition, product recommendations, medical diagnosis, self-driving cars, game opponents.
2Expert systems
An expert system simulates the knowledge and decision-making of a human expert in one area, e.g. medical diagnosis, fault-finding in a car, or finding mineral deposits.
| Component | Role |
|---|---|
| knowledge base | a database of facts about the subject, gathered from human experts |
| rule base | a set of rules (IF … THEN) used to reason with the facts |
| inference engine | applies the rules in the rule base to the facts in the knowledge base and the user's answers, to reach a conclusion; decides the next question to ask |
| interface | lets the user enter data and answer questions, and shows the conclusions, often with an explanation and a probability |
- The interface asks the user a series of questions (often yes/no).
- The inference engine searches the knowledge base and applies the rules in the rule base to the answers.
- It decides which question to ask next, until it can reach a conclusion.
- The conclusion (diagnosis or advice) is shown, with an explanation of how it was reached.
| Advantages | Disadvantages |
|---|---|
| consistent answers; does not forget or tire | no common sense or emotion; results only as good as the knowledge base |
| available any time; can hold the knowledge of many experts | expensive to set up and keep up to date |
| quick | users may trust it too much |
3Machine learning
When a program has the ability to automatically adapt its own processes and/or data.
Instead of being given every rule by a programmer, a machine learning program is given a large amount of data (e.g. thousands of labelled photos). It finds patterns, makes predictions, and uses feedback on whether they were right to change its own processes, so it improves over time. Examples: spam filters that learn which emails are spam, recommendation systems, face and speech recognition.
4Exam focus
Key words: artificial intelligence; reasoning; learning; expert system; knowledge base; rule base; inference engine; interface; machine learning.
Quick self-check
Answer in your head first, then reveal.
Simulation of intelligent behaviour by computers (1).
Knowledge base (1); rule base (1); inference engine (1); interface (1).
Interface asks the user questions (1); inference engine applies rules from the rule base (1); to facts in the knowledge base (1); conclusion shown with explanation (1).
The program can automatically adapt (1); its own processes and/or data (1).
