What Are Knowledge Based Systems?

Knowledge based systems are a class of artificial intelligence software in which domain-specific knowledge and the control mechanisms that apply it are explicitly separated into distinct components. This architectural separation distinguishes them from conventional programs, in which algorithmic logic and data are interwoven. The core idea is that a knowledge base containing facts, rules, and heuristics can be queried and reasoned over by a general-purpose inference engine, allowing the system to answer questions or recommend actions in ways that approximate the judgment of a human expert.

The field emerged from artificial intelligence research in the 1960s and 1970s, drawing on formal logic, cognitive science, and rule-based programming. Early landmark systems such as MYCIN (medical diagnosis) and DENDRAL (chemical structure identification) demonstrated that encoding domain expertise in explicit symbolic form could yield practical results in specialized problem-solving tasks.

Knowledge Representation

A knowledge base stores domain knowledge in a structured, machine-interpretable form. The most common representations include production rules (if-then condition-action pairs), frames (object-like structures with slots for attributes and values), semantic networks, and formal ontologies. As described in research on knowledge representation and inference in AI, the choice of representation directly shapes what kinds of queries the system can answer and how efficiently it can reason. Factual knowledge covers relationships and properties established in the domain; heuristic knowledge encodes experiential rules of thumb that practitioners apply when formal methods are insufficient.

Inference and Reasoning

The inference engine is the component that processes queries against the knowledge base. It applies two primary reasoning strategies. Forward chaining (data-driven inference) begins from known facts, matches them against rule conditions, and fires rules whose preconditions are satisfied, generating new facts until a goal is reached or no further rules apply. Backward chaining (goal-driven inference) starts from a desired conclusion and works backward through rule chains to identify what facts must be established, querying the user or database only when required. Conflict resolution strategies govern which rule to apply when multiple rules are simultaneously eligible. The Taylor & Francis knowledge engineering reference situates these mechanisms within the broader structure of AI reasoning research.

Decision Support Systems and Deductive Databases

Knowledge based systems underpin a range of decision support applications. A decision support system combines a knowledge base with data access and user interaction to assist human decision-makers rather than replace them, surfacing relevant rules, precedents, and recommendations on demand. Deductive databases extend relational database technology with logical inference, allowing users to query not just stored data but also facts that can be derived from that data through defined rules. This extension is particularly valuable when implicit relationships among entities need to be surfaced without pre-materializing every possible derived fact. The ACM Digital Library contains foundational work on both decision support integration and deductive database query languages, including Datalog-family formalisms.

Applications

Knowledge based systems have applications across many industries, including:

• Clinical decision support and medical diagnosis assistance
• Legal reasoning and regulatory compliance checking
• Process fault diagnosis in industrial manufacturing and control
• Financial risk assessment and credit evaluation
• Configuration management in complex engineering systems
• Scientific research assistance through automated hypothesis exploration