logo

Unit 3: Types of expert systems

Learning outcomes

By the end of this unit you should:

  • be able to differentiate between expert and non-expert systems
  • know the main categories of expert systems
  • understand 8 ways to represent expert systems visually
Rubik

Activity 1: Review

Work in pairs. Without reading your notes or the course website, explain the different types of knowledge.

Activity 2: Expert systems

Read the following passage to answer these questions

  1. What is an expert system?
  2. What are the main types of expert systems?

Expert systems are a branch of artificial intelligence that uses specialized knowledge to solve problems in a specific domain. These systems are capable of making decisions similar to a human expert. The information in an expert system is stored in the form of a knowledge base, and it uses a set of rules to interpret this information.

The first type is rule-based systems. These systems use a set of if-then rules to solve problems. For example, a medical diagnosis system might have a rule like "if the patient has a fever and a cough, then they may have the flu". The system applies these rules to the information it has to make a decision.

The second type is frame-based systems. These systems use frames, which are data structures for representing a stereotyped situation. Each frame has a number of slots, or attributes, which can contain values. For instance, a frame representing a car might have slots for color, make, model, and year.

The third type is fuzzy systems. These systems are designed to handle uncertainty and ambiguity. Instead of using exact values, fuzzy systems use degrees of truth, which are represented as values between 0 and 1. For example, a weather prediction system might say there's a 0.7 chance of rain tomorrow, instead of making a definite prediction.

Work in pairs. Discuss your answers to the two questions.

Activity 3: Comprehension check

Work in pairs. Find the answers to the following questions in the text in Activity 2. When you both agree on the answer (and its location in the text), move on to the next question.

An expert system is a branch of artificial intelligence that uses specialized knowledge to solve problems in a specific domain.

The knowledge base in an expert system stores the information that the system uses to make decisions.

A rule-based system is a type of expert system that uses a set of if-then rules to solve problems.

An example of a rule in a rule-based system could be judging a patient has flu based on the symptoms of a fever and a cough.

A frame-based system is a type of expert system that uses frames, or data structures, to represent stereotyped situations.

A frame in a frame-based system representing a car might have slots for colour, make (e.g. Subaru), model (e.g. Outback), and year.

A fuzzy system is a type of expert system designed to handle uncertainty and ambiguity.

Fuzzy systems represent degrees of truth as values between 0 and 1.

A weather prediction system might say there is a 0.7 chance of rain tomorrow, instead of making a definite prediction.

A fuzzy system might be best for a situation where there is a lot of uncertainty, as it is designed to handle such scenarios.

Check the answers by clicking on the question.

Activity 4: Criteria for an expert system

Read the criteria that are used to evaluate whether a system is an expert system.

A system may be categorized as an expert system if the following criteria are fulfilled:

  1. An expert system should have in-depth knowledge in a specific domain.
  2. It should have a knowledge base.
  3. An expert system should have an inference engine, i.e. a way to apply its knowledge to a specific situation.
  4. It can explain its reasoning process.
  5. Expert systems need a user interface, e.g text or GUI.
  6. Its performance should be comparable to that of a human expert in its field

Activity 5: Differentiating between expert and non-expert systems

Work in pairs. Decide whether the systems listed below are expert or non-expert systems.

Grammarly is a rule-based expert system. Grammarly has a set of rules related to English grammar, punctuation, spelling, and stylistic norms. When a student types a sentence, Grammarly compares the sentence to its rule set.

Netflix is a frame-based expert system. Netflix uses frame-based systems to categorize movies and TV shows into different frames or classes (for example, action movies, romantic comedies, documentaries, etc.). Each frame has a number of attributes like director, actors, length, viewer ratings, etc. When a student watches a movie, Netflix fills in the relevant frame's slots with the movie's attributes. By comparing this frame to other frames, Netflix can make recommendations based on similarities.

ChatGPT is not an expert system. ChatGPT is a general-purpose conversational AI that can generate responses across a wide range of topics, not just a single narrow domain. It does not provide expert advice. Its responses are based on patterns it has learned, and it can sometimes generate incorrect or nonsensical responses.

Google and other search engines might be confused for expert systems because they can retrieve specific information from a vast database in response to user queries. However, search engines primarily function by matching keywords and employing complex algorithms to rank results, rather than using structured, domain-specific knowledge to provide expert decisions.

Weather prediction apps are examples of fuzzy expert systems. Weather forecasting is an uncertain process. Instead of giving absolute predictions, these systems provide probabilities.

The Roomba is not an expert system. It is designed to carry out a specific set of actions based on its inputs (sensor data), but it doesn't solve complex problems in a specific knowledge domain like an expert system would.

Activity 6: Designing expert system: visuals

Read the 8 most frequent ways to visualize expert systems.

  1. Decision tree: A decision tree is a hierarchical structure that uses nodes to represent conditions or decisions and branches to represent possible outcomes or actions. It is an intuitive way to represent a series of decision points and the corresponding actions to be taken.
  2. Flowchart: A flowchart is a graphical representation of a process or system, using various shapes and arrows to illustrate the flow of steps or decisions. It can effectively capture the logical flow of rules and actions in an expert system.
  3. Rule table: A rule table is a tabular representation of rules in which each row represents a rule and each column represents a condition or action. This format is particularly useful when there are multiple rules with similar structures but varying conditions and actions.
  4. Rule network: A rule network visualizes the relationships between rules using nodes and connections. It shows how rules interact and depend on each other, providing a more holistic view of the expert system's knowledge base.
  5. Rule graph: A rule graph represents rules as nodes and depicts the relationships between them using directed edges. It emphasizes the dependencies and connections between rules in a more visual and interconnected manner.
  6. State transition diagram: A state transition diagram illustrates the different states and transitions of a system based on specific conditions. It is particularly useful when modeling systems with dynamic behavior or complex rule interactions.
  7. Natural Language Processing (NLP) Visualization:: For complex rule-based expert systems that involve natural language processing or text analysis, visualizations can be used to represent the flow of text and linguistic processing steps. This type of visualization can help understand how rules and conditions are applied to text data.
  8. UML class diagrams: While primarily used for object-oriented software modeling, a UML (Unified Modeling Language) class diagram can be adapted to represent rule-based expert systems. It can depict the classes, attributes, methods, and relationships between rules and rule sets.

Activity 7: Visualing expert system: Practice

Work in pairs or threes. Use TWO different ways to represent one of the systems listed below. Both ways must show exactly the same decision process. Submit your work on ELMS.

  1. Tenses: The input of this system is any sentence about future plans and predictions with the verb shown in brackets, e.g. Tomorrow, it (rain) in the morning. The system should select whether to use will future or going-to future. The output is a grammatically-correct sentence using either will future or going-to future.
  2. Stressed syllables: The input of this system is any word. The system should identify the number of the syllables and identify which syllable carries the primary stress. The output of the system should show which syllable to stress.
  3. Adjective order: The input of this system is an series of adjectives. The system should identify the type of adjective. The output of the system should which the most frequent (i.g. natural or unmarked) order of the adjectives.

Review

Can you:

  1. do this
  2. do that
  3. and do something else.

If you do not, make sure that you do before your next class.

Running count: 38 of 38 concepts covered so far.