Frequently Asked Questions

A diagnostic expert system is a knowledge-based software system that uses a collection of rules and facts — encoding the knowledge of human experts — to reason about symptoms or observations and reach a diagnosis or recommendation. "Diagnostic" refers to systems that determine the cause of a problem from observed evidence. See the Expert Systems chapter for the full architecture.

Expert systems encode knowledge explicitly as human-readable if–then rules, built through collaboration with domain experts. A machine learning system like a neural network learns its rules implicitly from training data — the rules are encoded as numeric weights, not human-readable logic. Expert systems are easier to inspect and explain; machine learning systems can discover patterns automatically but are harder to interpret.

The knowledge base is the component of an expert system that stores all domain knowledge — the facts (known truths) and rules (if condition then conclusion) that the system uses to reason. Building and maintaining the knowledge base is the central task in developing an expert system, typically requiring collaboration with domain experts.

An inference net (or inference network) is a graphical representation of the reasoning process in a diagnostic expert system. Nodes represent hypotheses (possible diagnoses) or observable evidence; arcs represent causal or evidential relationships between them. The inference engine traverses the net to determine which hypotheses are supported by the observed evidence. See the Inference Net chapter and the worked example.

Forward chaining starts from known facts and applies rules to derive new conclusions (data-driven reasoning). Backward chaining starts from a goal — a hypothesis to confirm or deny — and works backwards to find supporting evidence (goal-driven reasoning). Diagnostic expert systems typically use backward chaining: they start with a potential diagnosis and look for evidence that supports or refutes it.

Yes. Expert systems are used in medical diagnosis, technical troubleshooting, legal reasoning, financial advisory, and configuration systems. Modern AI has complemented rather than replaced them — hybrid systems combine rule-based reasoning with machine learning for greater robustness and coverage.

The text and images in this tutorial are published under the Creative Commons Attribution–ShareAlike 4.0 licence (CC BY-SA 4.0). You are free to share and adapt them, provided you give appropriate credit and distribute any adaptations under the same licence.

The interactive JavaScript demonstrations may only be used as part of this website; any other use requires prior written permission. Contact: marek@obitko.com.