An expert system is an example of a knowledge-based system. Expert systems were the first commercial systems to use a knowledge-based architecture. A knowledge-based system is essentially composed of two sub-systems: the knowledge base and the inference engine. (https://en.wikipedia.org/wiki/Expert_system)
The ideal representation for a knowledge-base is an object model (often called an ontology in AI literature) with classes, subclasses, and instances. (Cf. https://en.wikipedia.org/wiki/Knowledge_base)
Components of ES:
- This is a repository of information, facts, and rules relevant to a particular field. Think of it as the expert's knowledge, stored in a structured way.
- This component uses logical rules to process information from the knowledge base and derive new conclusions or solutions to problems. It acts like the expert's reasoning process.
- Allows users to interact with the system, input information about a problem, and receive advice or solutions.
- ES can give choices like Recommendation system.
- Instead of a fixed set of rules, the knowledge base can be a trained machine learning model. The inference process can be no longer just logical deduction. The ML model itself acts as the inference engine by taking new data as input and producing a prediction or classification as output. It uses the patterns it learned from the data to "reason" and arrive at a conclusion.
- One of the main drawbacks of many ML models is their lack of interpretability (the "black box" problem).
By integrating an ML model into a larger expert system, you can use the rule-based component to provide a human-readable explanation for the ML model's output. For instance, an expert system might use an ML model to detect a financial anomaly, then use its rule-based logic to explain that the anomaly is due to a transaction in a foreign country from an unknown user, which aligns with a fraud-detection rule.
