Artificial Intelligence

Learning Outcomes

Upon successful completion of this course, students should be able to

• Explain fundamental concepts including agents, problem & state/action/ spaces, problem solving via search as a model of thinking, heuristics, knowledge representation and reasoning using logic.
• Select algorithms for problem solving based on problems characteristics, and characteristics of problem & state/action spaces.
• Evaluate usefulness, advantages and limitations of various algorithms and methods towards increasing computational effectiveness of problem solving
• Modelling problems as constraint satisfaction problems, or as problems in logic.

Course Contents

• Introduction to artificial intelligence, goals, advances, prospects, limitations, and basic notions regarding agents and problem solving.
• Blind search algorithms
• Informed search algorithms and proofs of finding optimal solutions
• Heuristic functions and their construction and selection
• Local search using hill climbing, simulated annealing, local beam search, genetic algorithms.
• Constraint problem solving: depth first, depth first with various forms of backtracking, forward checking, arc consistence, maintaining arc consistency, min conflicts.
• Knowledge representation and reasoning in logic: Propositional logic, entailment, , resolution, satisfiability, DPLL, local search for satisfiability.

Recommended Readings

• Stuart Russel and Peter Norvig. Artificial Intelligenc­e: A Μodern Approach, Prentice Hall, 2nd edition (2003). http://aima.cs.berkeley.edu/. Το βιβλίο έχει εκδοθεί στα Ελληνικά από τις εκδόσεις Κλειδάριθμος με τον τίτλο «Τεχνητή Νοημοσύνη: Μια σύγχρονη προσέγγιση».http://aima.uom.gr/.
• Ι. Βλαχάβα, Π. Κεφαλά, Ν. Βασιλειάδη, Φ. Κόκκορα και Η. Σακελαρίου. Τεχνητή Νοημοσύνη. Εκδοτικός οίκος «Β. Γκιούρδας Εκδοτική – Μονοπρόσωπη ΕΠΕ».http://aibook.csd.auth.gr.