Symbolic AI
Symbolic AI represents knowledge with explicit symbols, rules or logical statements and uses procedures to reason over them. Examples include rule-based inference, planning and theorem proving.
A delivery system could store the rule “If an order contains a refrigerated item, use cold-chain shipping.” Given an order marked as refrigerated, an inference procedure derives the shipping requirement. The result follows from represented facts and a written rule, rather than from a statistical pattern learned from labeled orders.
That makes the reasoning path inspectable, but it does not establish that the facts or rule are correct. An incorrectly tagged item can still produce a wrong decision. Real applications need an owner for rules, a way to resolve conflicts and tests for the conditions that matter.
Symbolic AI is broader than an expert system and narrower than the informal label traditional AI. A system can also combine symbolic constraints with learned predictions—for example, a model estimates delivery demand while rules enforce known shipping requirements.
Sources
- Poole and Mackworth: Propositions and Inference — Introduces propositional representations and reasoning over explicit knowledge.
Go deeper
- UC Berkeley CS188: Logic course
Work through logical knowledge representation and inference in an AI course.