Information Technology • Expert Systems

Expert System Builder

Rule base

    Knowledge base: what is known so far

    User interface

    Expert System Builder — Explanation

    How to use it. Choose a system in the title bar. Ask me runs a consultation: the inference engine works backwards from each possible conclusion and asks you only the questions it needs (press Why? at any question). From facts lets you tick what you know and watches forward chaining fire the rules. Edit rules lets you change the facts and rules, or add your own; the other modes then use your version. The strip at the top shows the parts of an expert system, lit up as each one works: click one to read about it.

    The parts

    A knowledge base editor lets the experts (or a knowledge engineer) add and change facts and rules: that is what Edit rules is.

    Chaining

    Backward chaining is goal-driven: it picks a possible conclusion and works back through the rules to the facts it needs, asking about them. It suits diagnosis. Forward chaining is data-driven: it starts with the facts it is given and fires every rule whose IF part is true, adding each conclusion as a new fact, in cycles, until nothing new can be concluded. Either way, one rule's conclusion can be another rule's condition: that is the chain.

    Uses and limitations

    Uses: medical diagnosis, car and machine fault-finding, plant and animal identification, mineral prospecting, financial and insurance decisions, legal advice, chess. Advantages: consistent answers, available at any time and in many places at once, quick, can explain its reasoning, and keeps the knowledge of experts who retire. Limitations: no common sense, judgement or empathy; only as good as its rules (and wrong outside them); expensive to build and to keep up to date; and people may trust it too much. It can't learn on its own.

    Common exam mistakes: mixing up the knowledge base (facts) and the inference engine (reasoning); describing forward chaining as starting from the goal; and listing "it can think for itself" as an advantage.

    Objective: Cambridge International AS & A Level Information Technology (9626), expert systems: their components, how they are created and used, forward and backward chaining, and their uses and limitations. The two systems here are teaching examples, not real advice.

    Where this fits

    • Cambridge: Cambridge IGCSE Information and Communication Technology (0417); Cambridge IGCSE Computer Science (0478); Cambridge A Level Information Technology (9626); Cambridge AS Level Information Technology (9626) Goes beyond Cambridge IGCSE Information and Communication Technology (0417): Forward and backward chaining go beyond 0417. Goes beyond Cambridge IGCSE Computer Science (0478): Forward and backward chaining go beyond 0478.
    • Pearson Edexcel International: Edexcel International A Level Information Technology

    Expert System Builder — Key Terms

    Key concepts in English, with te reo Māori, Chinese (Simplified) and Korean.

    EnglishTe reo Māori中文(简体)한국어What it means on this page
    Expert Systemno attested term专家系统 (zhuānjiā xìtǒng)전문가 시스템 (jeonmunga siseutem)A computer system that uses a knowledge base and rules gathered from human experts to give advice or reach conclusions in one subject.
    Knowledge Baseno attested term知识库 (zhīshikù)지식 베이스 (jisik beiseu)The part of an expert system that stores the facts about its subject.
    Rule Baseno attested term规则库 (guīzékù)규칙 베이스 (gyuchik beiseu)The part of an expert system that stores the IF … THEN rules linking facts to conclusions.
    Inference Engineno attested term推理机 (tuīlǐjī)추론 엔진 (churon enjin)The reasoning part of an expert system: it applies the rules to the facts to reach conclusions and decides what to ask next.
    User Interfaceno attested term用户界面 (yònghù jièmiàn)사용자 인터페이스 (sayongja inteopeiseu)The part of a system that people interact with: here, it asks the questions and shows the conclusions and advice.
    Explanation Systemno attested term解释系统 (jiěshì xìtǒng)설명 시스템 (seolmyeong siseutem)The part of an expert system that shows how it reached a conclusion and why it is asking a question.
    Knowledge Base Editorno attested term知识库编辑器 (zhīshikù biānjíqì)지식 베이스 편집기 (jisik beiseu pyeonjipgi)The tool experts and knowledge engineers use to add, change and delete the facts and rules of an expert system.
    Forward Chainingno attested term正向推理 (zhèngxiàng tuīlǐ)전방 추론 (jeonbang churon)Reasoning from the known facts: firing every rule whose IF part is true and adding its conclusion, until nothing new can be concluded.
    Backward Chainingno attested term反向推理 (fǎnxiàng tuīlǐ)후방 추론 (hubang churon)Reasoning from a possible conclusion (a goal) back to the facts that would support it, asking for those facts.

    On the te reo Māori column. Terms marked as gaps have no attested equivalent in the sources checked — Karaitiana Taiuru's Dictionary of Māori Computer and Social Media Terms, Paekupu, the Reserve Bank's te reo financial glossary, NZQA and Te Aka. No coinage is printed as though it were established; where a class needs one, commission it from Te Taura Whiri i te Reo Māori and credit the translator. Te reo Māori is not italicised and takes no plural "s".