Choose between an expert system and machine learning
Match the syllabus AI approach to a task and explain its evidence and limitations.
Exam scope and assessment notes
This sequence follows Cambridge IGCSE Computer Science 0984, 2026–2028 syllabus Version 4. It covers subject content sections 6.1–6.3 only. The separately published 2029–2031 syllabus is not mapped here. Practice is original, self-checked and does not award an official Cambridge grade. Full timed papers and human subject-teacher review remain outstanding. Cambridge states that 0984 has the same 2026–2028 subject content as 0478, with 9–1 grading. In June and November, Cambridge lists 0984 in Administrative zone 3 and for centres in Oman and the UAE. Confirm availability for your centre and exam series with Cambridge International. The separately published 2029–2031 syllabus changes to Python 3 only and removes pseudocode; this 2026–2028 route does not cover those changes.
Before you begin
- No programming language is required. Be comfortable with place value and multiplication; each lesson introduces the computer-science notation it uses.
Work on paper. Try each question before revealing help, and explain why your method works.
Quick readiness check
Which approach learns patterns from examples?
What you will learn
- Compare rule-based expert systems with adaptive machine learning
- Select an approach for a given task
- Justify human review and testing
Different AI approaches use information differently
An expert system stores facts in a knowledge base and rules in a rule base; its inference engine applies those rules to supplied information. It is suitable when expert knowledge can be expressed clearly as conditions.
Apply and check the rule
Machine learning adapts a program's processes or data from examples. It can be useful when patterns are difficult to write as fixed rules and suitable training data exists. Both approaches can fail when information is incomplete or the situation differs from their design, so evaluate outputs and retain human oversight for important decisions.
Pause and explain
A task has stable, clearly stated conditions. Which approach may be easiest to explain?
Worked example
A school wants to check whether a submitted form contains all required fields. Which approach may be simplest?
Show the worked solution
- The requirements are explicit and stable.
- They can be written as clear rules.
- A rule-based expert system is more suitable than learning from many examples.
Answer A rule-based approach can check each required field because the conditions are explicit.
Try a topic test
Choose Easy, Normal or Difficult. Save your answers and open worked review after finishing.
Test this topic →From guided practice to a new situation
Write your answer and reasoning first. Use a hint only if you are stuck. The model solution is for self-checking; your written work is not automatically marked.
Choose an expert system or machine learning to classify unusual images when no clear rule can be written but many labelled examples exist.
Give me a hint
Select based on how the task can be represented.
Compare my reasoning
- The task has no simple explicit rule.
- There are many labelled examples.
- Machine learning can adapt a classification process from the examples.
Machine learning is a plausible choice because labelled examples are available and the pattern is difficult to express as fixed rules.
Look for these in your work
- Task feature
- Data available
- Approach justified
Choose a suitable approach for a fixed troubleshooting guide with expert-written conditions, and name its main components.
Give me a hint
Use explicit facts and rules.
Compare my reasoning
- The conditions are clearly stated.
- Store facts in a knowledge base and conditions in a rule base.
- An inference engine applies rules and an interface presents the advice.
An expert system fits: it uses a knowledge base, rule base, inference engine and interface.
Look for these in your work
- Approach selected
- Components
- Task match
A team proposes machine learning for a high-impact eligibility decision. State two checks needed before use and explain why oversight remains important.
Give me a hint
Test coverage and consequences of errors.
Compare my reasoning
- Evaluate accuracy on representative cases not used in training.
- Check for uneven errors across relevant groups and monitor changes over time.
- Keep a qualified person able to review or challenge consequential decisions.
Test on representative unseen cases, inspect errors across groups and monitor performance. Human oversight matters because learned patterns can be wrong or biased and decisions have real consequences.
Look for these in your work
- Unseen evaluation
- Fairness check
- Human accountability
Common mistakes
- Claiming machine learning is always more accurate
- Using an expert system when rules cannot represent the changing task
What can you explain now?
Which approach uses a knowledge base and rule base?
Compare with the explanation
Expert system
Its inference engine applies stored rules to stored or supplied facts.
After trying it yourself, choose your next review. This is your self-assessment.
Your review choice appears on Today. Sign in to sync it across devices.
Make it stick
When might an expert system fit better than machine learning, and when might the reverse be true?
If you needed the solution, close it and solve the problem again from a blank page. A correct answer is stronger when you can explain the reason for each step.