CSEN 907 Knowledge Representation and Reasoning (Elective for MET)

## Course Information

### Abstract

• Knowledge represenation and reasoning (KRR) is considered to be the sub-filed of artificial intelligence concerned with the representation of information in computers in ways that allow computers to draw reasonable conclusions from them. Representation is mostly declarative, using a formal language with clear syntax and semantics, and inference rules that define what conclusions are reasonable---in other words, a logic. Classical logics, such as propositional and first-order logics, were primarily designed to lay the foundations of mathematics. But, when it comes to commonsense KRR, such logics need to be extended/modified to effectively and efficiently capture human-like reasoning modes. This course covers the general principles of knowledge representation and reasoning. Students who pass this course will have acquired the fundamentals needed to pursue research in any of the specialized KRR areas.

### Outline

• What is "knowledge", "representation", and "reasoning"?
• Propositional logic
• Reasoning in propositional logic
• First-order logic
• Reasoning in first-order logic
• Formalizing commonsense reasoning
• Modal logic
• Non-monotonic logic
• The logic of time
• Problems in non-monotonic temporal reasoning
• Belief change and reason maintenance
• The logic of causality

### Objectives

• After passing this course, students should be able to do the following.

1-Argue for and against the definition of knowledge as justified true belief.
2-Identify different types of reasonig.
3-Prove properties of the syntax of propositional logic.
4-Prove properties of the semantics of propositional logic.
5-Compute the interpretation of a propositional WFF given a truth assignment.
6-Determine whether a propositional WFF is valid (contradictory, satisfiable).
7-Determine whether a set of propositional WFFs logically imply a propositional WFF.
8-Determine whether two propositional WFFs are logically equivalent.
9-Identify valid propsitional arguments.
10-Apply a tableau method to determine the validity of a propositional argument.
11-Construct propositional derivations using natural deduction.
12-Demonstrate 3--11 for first-order logic.
13-Construct detivations using resolution refutation.
14-Unify two expressions.
15-Translate English sentences into first-order logic WFFs.
16-Construct domain theories for simple domains.
17-Identify the necessity of domain closure axioms and unique-names axioms.
18-Demonstrate 3--9 and 11 for propositional modal logic.
19-Demonstrate 3--9 and 11 for first-order modal logic.
20-Distinguish between de re and de discto readings of sentences.
21-Construct possible worlds semantics for systems with and without the Barcan formula.
22-Correctly represent knowledge involing equality in first-order modal logic.
23-Compute entailments given the closed-world assumption and the general closed world assumption.
24-Compute the circumscription of a theory.
25-Compute extensions of default theories.
26-Identify Allen's interval relations.
27-Represent temporal reports using tense logic and Allen's interval logic.
28-Identify events, processes, and states.
29-Analyze theories suffering from the frame problem, the Yale shooting problem, and the ramification problem.

### Textbooks

• Knowledge Representation and Reasoning
Ronald Brachman and Hector Levesque.
Morgan Kaufmann Publications. ISBN-10 1-55860-932-6