- Learn
- Artificial Intelligence
- Knowledge, Logic and Planning
AI-M02
Knowledge, Logic and Planning
How a system can store what it knows and reason with it: facts and rules, logic, expert systems, knowledge graphs and planners.
Search finds paths through a space someone else defined. This module asks where that knowledge comes from and how a machine can reason with it. You will write facts and rules down explicitly, check what follows from them with propositional and first-order logic, and see how the same ideas powered the expert systems of the 1970s and 1980s, and why those systems broke.
Then the ideas move into the present: knowledge graphs that answer questions in search engines, rule engines that still make auditable decisions in banking and compliance, and planners that turn a goal into a sequence of actions. Language-model agents that call tools and plan multi-step work lean on every one of these ideas, so this is the background that makes them easier to understand.
Lessons
- AI-02.01 Knowledge representation Knowledge representation is about storing what a system knows in a form it can reason with: facts, rules, objects and their relationships. Beginner 10 min
- AI-02.02 Propositional logic Propositional logic combines true-or-false statements with and, or, not and implies. Intermediate 14 min
- AI-02.03 First-order logic First-order logic adds objects, relations and quantifiers such as for all and there exists, so you can state general facts like every human is mortal. Intermediate 14 min
- AI-02.04 Expert systems and rule engines Expert systems encode specialist knowledge as if-then rules and use an inference engine to reach conclusions, famously in medical diagnosis research in the 1970s. Beginner 10 min
- AI-02.05 Knowledge graphs and ontologies A knowledge graph stores facts as triples, subject, relation and object, such as Paris is the capital of France. Intermediate 12 min
- AI-02.06 Classical planning Classical planning finds a sequence of actions that turns a starting state into a goal state, where each action has preconditions and effects, described in languages like STRIPS and PDDL. Advanced 12 min