Artificial Intelligence
Search — uninformed (BFS, DFS, uniform-cost, iterative deepening), informed (greedy best-first, A*, heuristics) and adversarial (minimax, alpha–beta); logic — propositional and predicate (first-order) logic, inference and resolution; and reasoning under uncertainty — conditional independence and Bayesian networks, exact inference by variable elimination, and approximate inference by sampling.
The AI section covers three classical pillars: search (how an agent finds a sequence of actions), logic (how it represents and reasons with certain knowledge) and probabilistic reasoning (how it reasons with uncertain knowledge). Third-party analyses put it at about 11 marks in 2024 and only 3 in 2025, so the weight varies — but the questions are usually quick marks if you can trace an algorithm by hand.
What the syllabus lists: search — informed, uninformed, adversarial; logic — propositional, predicate; reasoning under uncertainty — conditional independence representation, exact inference through variable elimination, and approximate inference through sampling.