Expert Systems in Artificial Intelligence

Podcast

Play an AI-generated podcast conversation about this lesson

Questions and Answers

What is the primary function of a rule in artificial intelligence?

To create algorithms for data processing.

To outline a procedure for problem-solving. (correct)

To represent knowledge in a complex manner.

To convey information as a single statement.

What do the keywords AND and OR accomplish in a rule?

They allow multiple antecedents in a rule. (correct)

They restrict the rule to a single antecedent.

They define the consequent of the rule.

They specify the type of action to be taken.

How is the relationship between an object and its value established in a rule's antecedent?

By using mathematical operators only.

Through a conditional statement.

With the help of an operator linking them. (correct)

By assigning a numerical value to the object.

Which part of the rule structure provides the condition under which an action is taken?

The antecedent. Signup and view all the answers

Which of the following statements about expert systems is accurate?

Expert systems can express knowledge through rules. Signup and view all the answers

What is the primary role of a knowledge engineer in the development of an expert system?

To interview domain experts and determine reasoning methods Signup and view all the answers

Which programming languages are essential for a programmer working on an expert system?

C, Basic, LISP Signup and view all the answers

What is a critical responsibility of the project manager in an expert system development team?

To guarantee all project milestones are met Signup and view all the answers

In rule-based expert systems, where is problem-specific information generally stored?

Short-term memory Signup and view all the answers

Which of the following best describes the end-user's role in the expert system development process?

Provides input on user interface design Signup and view all the answers

What foundational model did Newell and Simon propose for rule-based expert systems?

Production system model Signup and view all the answers

What aspect of user experience is vital for the success of an expert system?

The design of the user interface Signup and view all the answers

Which reasoning technique starts with a hypothetical solution to find evidence?

Backward chaining Signup and view all the answers

What does forward chaining primarily focus on?

Gathering information to infer conclusions Signup and view all the answers

In backward chaining, if the current goal cannot be proved, what does the inference engine do?

It stacks the current rule and sets a new subgoal Signup and view all the answers

Which statement correctly describes the inefficiency of forward chaining?

It often proves multiple unrelated goals at once Signup and view all the answers

Which of the following describes a characteristic of backward chaining?

It proves IF parts rather than THEN actions Signup and view all the answers

What is the role of the knowledge base in backward chaining?

To hold rules needed to prove the current subgoal Signup and view all the answers

What happens when two rules have the same IF part but different THEN actions?

They create ambiguity in rule execution Signup and view all the answers

When should one prefer forward chaining over backward chaining?

When starting from existing facts to infer new conclusions Signup and view all the answers

Which rule among the following represents a potential conflict?

IF traffic light is red THEN action is stop Signup and view all the answers

What structure do rules in a rule-based expert system follow?

IF (condition) THEN (action) Signup and view all the answers

What role does the inference engine play in an expert system?

It matches rules with the facts to reach a solution. Signup and view all the answers

Which of the following is NOT a characteristic of an expert system?

Speed of reaching a solution is irrelevant. Signup and view all the answers

How does an expert system explain its conclusions?

By showing the logical reasoning and justifications used. Signup and view all the answers

What type of reasoning do expert systems typically employ when solving problems?

Symbolic reasoning representing different types of knowledge. Signup and view all the answers

What is the primary importance of speed in an expert system?

Timeliness can impact outcomes in critical situations. Signup and view all the answers

What is the function of the database in a rule-based expert system?

To provide facts that match the IF conditions of the rules. Signup and view all the answers

Which of the following best describes heuristics in expert systems?

Strategies that reduce the problem search area. Signup and view all the answers

What aspect is critical for user satisfaction in relation to expert systems?

The accuracy of the solutions provided. Signup and view all the answers

Study Notes

Expert Systems Overview

An expert system is a computer program designed to hold the accumulated knowledge of one or more domain experts.
Expert systems are a branch of Artificial Intelligence that aim to mimic human experts.
Expert systems can support decision-makers or entirely replace them.
They are the most widely applied and commercially successful Artificial Intelligence technology.

Components of an Expert System

The knowledge base is a collection of facts and rules that describe the knowledge about the problem domain.
The inference engine is the part of the system that chooses which facts and rules to apply when trying to solve the user’s query.
The user interface is the part of the system that takes in the user's query and passes it to the inference engine.

Why use Expert Systems?

Experts are not always available, and an expert system can be used anywhere, any time.
Human experts are often not 100% reliable or consistent.
Experts may not be good at explaining their decisions.
Expert systems are often cost-effective.

Applications of Expert Systems

DENDRAL: Used to identify the structure of chemical compounds (first used in 1965).
LITHIAN: Provides advice to archaeologists examining stone tools.

Problems with Expert Systems

Limited domain: systems are restricted to a specific area of expertise.
Lack of up-to-date knowledge: Expert systems may not always have the most recent information and do not learn.
Lack of common sense: systems do not have the ability to use common sense or apply general knowledge.
Requires experts for setup and maintenance: expert knowledge is needed for creation, development, testing, and maintenance.

Legal and Ethical Issues

Who is responsible if the advice is wrong? This is a complex question with no single answer. Possible responsibilities include the user, the domain expert, the knowledge engineer, the programmer, and the company selling the software.

Expert Systems on the Web

Included are web links to demonstrations of expert systems. (Specific web addresses are not included because they may not be relevant long-term and are not part of the core content of the expert system.)

Introduction to Knowledge

Knowledge is a theoretical or practical understanding of a subject or a domain.
Knowledge is also the sum of what is currently known, and apparently knowledge is power. Those who possess knowledge are called experts.
Anyone with deep knowledge (facts and rules) and experience in a particular domain can be considered a domain expert.

Rules for Problem Solving

The human mental process is complex and can't be fully represented as a simple algorithm.
However, experts can express their knowledge using rules for problem solving, such as IF a condition is met THEN an action takes place, e.g. IF the "traffic light" is green THEN the action is to go.

Rules as a Knowledge Representation Technique

A rule, in AI, is an IF-THEN statement that relates given facts or information in the IF part to an action in the THEN part.
A rule describes how to solve a problem.
Rules are typically easy to create and understand.
Any rule consists of an antecedent (IF part) and a consequent (THEN part).

Antecedent and Consequent

The antecedent of a rule incorporates two parts: an object and its value.
The operator links the object and its value.
Operators, like is, are, is not or are not, assign symbolic or numerical values to linguistic or numerical objects.
Expert systems can also use mathematical operations to represent the numerical value of an object.

Rules representing Relations, Recommendations, Directives, Strategies and Heuristics

(Various examples are provided. Specific examples are not included because their use case value is not part of the core expert system.)

Main Players in Development Team

The expert system development team consists of five key players: the domain expert, the knowledge engineer, the programmer, the project manager, and the end-user.
The success of the expert system depends on how well the team members work together and communicate.

Domain Expert

The domain expert is a knowledgeable person who is skilled in solving problems in a specific area/domain.
They have the greatest level of expertise in the domain.
They must be able to communicate their knowledge effectively.
They are also responsible for participating in the development and devoting significant time to making the project successful.

Knowledge Engineer

The knowledge engineer designs, builds, and tests expert systems.
They interview domain experts to understand how problems are solved.
They define reasoning methods, decide on knowledge representation, select development software or shells, program knowledge, and test/revise/integrate the system with the workplace.

Programmer

The programmer is responsible for the actual programming.
Their role is to describe knowledge using computer-understandable terms.
Typically skilled in symbolic programming languages like Lisp, Prolog, or OPS5 and have experience with expert systems, and programming languages such as C, Pascal, FORTRAN and Basic.

Project Manager

The project manager leads the expert system development team.
They ensure tasks are completed on time and deliverables are met.
They interact with all project members, including the expert, knowledge engineer, programmer and end-user.

End-User

The end-user uses the expert system.
The user not only needs to trust the system's performance, but also feel comfortable using it.
The expert system's design and its intuitive interface are crucial for end-user success.

Rule-Based Expert System Structure

The early model of rule-based expert systems used a production system model.
This is based on humans applying specific knowledge to solve a problem that uses rules.
Rules are stored in long-term memory and problem-specific facts are stored in short-term memory.

Knowledge Base

The knowledge base contains the domain expertise used for problem solving.
It is presented via rules in IF-THEN structure
(e.g., If 'condition', THEN 'action')
The expert system database is a set of facts that match the rules in the knowledge base.

Inference Engine

The inference engine carries out the reasoning process used by the expert system to derive a solution to the user's query.
It links the rules in the knowledge base to the facts in the database.

Explanation Facilities

The explanation facilities allow the user to ask the system how solutions are arrived at and why specific facts are needed.
Expertise in systems should be able to explain its reasoning.

Characteristics of an Expert System

Expert systems are built to function at the level of human experts in a specific domain.
High quality performance is the key characteristic; mistakes invalidate the high quality of output.
Speed is a key factor, but efficiency must be balanced with the need to produce useful results, such as in disaster recovery.

Expert System Use of Heuristics

Expert systems use heuristics to narrow down solutions.
A unique feature of expert systems is their explanation capability: systems are able to explain their reasoning and conclusions.
They utilize symbolic reasoning for the knowledge representation.

Expert Systems and Mistakes

Systems performing at the level of human experts are also able to make mistakes.
Experts are trusted often, even though they can make wrong judgements or decisions.
It is important to be careful, however, when using outputs from expert systems for critical decisions.

Knowledge Separation

In expert systems, knowledge and processing are separated into separate components (knowledge base and inference engine), unlike conventional programs.
This separation aids in understanding and modifying the code and makes development flexible.

Forward Chaining

Forward chaining, a data-driven reasoning technique, begins with known data and, using rules in the knowledge base, infers new facts and adds them to the database.
The process continues to add new facts until no additional rules can be applied.

Forward Chaining versus Backward Chaining

Forward chaining is suitable for gathering initial data and inferring new facts from it.
Backward chaining is suitable for a specific goal-driven hypothesis.

Backward Chaining

Backward chaining, a goal-driven technique, begins with a potential solution (hypothetical) and works backward to confirm it. Rules that address the solution are searched, and their conditions are checked to see if they match data in the database.
The system continues checking subgoals until a chain or complete verification is reached or until no more applicable rules can be found.

Methods of Choosing Between Forward and Backward Chaining

Choosing between forward and backward chaining depends on the starting point of the analysis.

Conflict Resolution

Conflict resolution is used when multiple rule antecedents in a rule base match the data.
It determines which rule takes precedence in firing by applying different strategies (prioritization of rules, specific versus general rules, or time-based prioritization).

Meta-knowledge

Meta-knowledge is knowledge about knowledge.
Metarules determine the strategy for using task-specific rules in an expert system.
Expert-supplied rules are typically given higher priority than those given by novices.

Advantages of Rule-Based Systems

Natural representation of knowledge: uses natural language for experts to explain the procedures.
Uniform structure of rules: this aids in their self-documentation.
Knowledge and processing are separate: systems are adaptable to a wider range of applications using the knowledge base and inference engine.
Dealing with uncertain knowledge: this function is typically part of the knowledge representation.

Disadvantages of Rule-Based Systems

Opaque rules: the complex interaction of several rules can make it hard to comprehend how the system arrives at a particular conclusion.
Inefficient search strategy, often slow for very large systems: this is a significant issue for some applications.
Inability to learn: expert systems may not be able to automatically alter the set of rules in the knowledge base, and further require the knowledge engineer to manage changes.

Studying That Suits You

Use AI to generate personalized quizzes and flashcards to suit your learning preferences.

Description

Test your knowledge on the principles and functions of rule-based expert systems in artificial intelligence. This quiz covers various aspects such as conditions, relationships, programming languages, and the roles in development teams. Evaluate your understanding of expert systems and the contributions from knowledge engineers and project managers.