COMP304: Knowledge Representation and Reasoning

Admin | Module Description | Lecture Notes

Admin

Lecturer:

Dr. Katie Atkinson.
Department of Computer Science,
Room 217,
Ashton Building,
Ashton St.

Email me at: katie [at] liverpool [dot] ac [dot] uk

Lecture time and locations:

• Tuesday 14.00, Ashton Lecture Theatre (ALT)
• Thursday 15.00, ALT
• Friday 10.00, ALT

Practical Class Information:

• Tuesday 15.00, ALT
• Friday 11.00, ALT

Practical classes will start in week 2 of the semester. Each person will attend one practical class: the allocations to the classes will be posted on VITAL before the classes commence.

Tutor:

Mr. Rolando Medellin.
Room 213,
Ashton Building,
Email: medellin [at] liverpool [dot] ac [dot] uk

Reading Material:

• M. Huth and M. Ryan.
  Logic in Computer Science: Modelling and reasoning about systems.
  Cambridge University Press, 2004.
  
  
• S. Russell and P. Norvig.
  Artificial Intelligence: A Modern Approach.
  Prentice Hall, 2003.
  
  

Additionally there are numerous other books in the library that deal with knowledge representation and logics for AI, and further pointers to additional reading may given in the lecture notes as the module proceeds.

[Back to top]

Module Description

Aims:

• To introduce the knowledge representation problem and logic as a unifying formalism.
• To study modal logics and their use.
• To introduce and demonstrate the use of model checking for modal logics.
• To provide an appreciation of modal logics that can be used to reason about knowledge and belief.
• To introduce and study non-monotonic logics and their use.
• To study methods for reasoning under uncertainty in agent systems.
  

Learning Outcomes:

At the end of the module you should:

• be able to explain and discuss the need for formal approaches to knowledge representation in artificial intelligence, and in particular, the value of logic as such an approach;
• be able to demonstrate knowledge of the basics of modal logic;
• be able to determine the truth/satisfiability of modal formulae;
• be able to perform modal logic model checking on simple examples;
• be able to model problems concerning agents' knowledge using epistemic logic;
• be able to demonstrate knowledge of the basics of non-monotonic logic and default reasoning;
• be able to model simple examples through the use of non-monotonic logic;
• be able to demonstrate knowledge of the basics of probability and decision theory, and their use in addressing problems in knowledge representation;
• be able to model simple problems involving uncertainty, using probability and decision theory.
  

Assessment:

• There is a continual assessment component for the module that counts for 20% of the final mark. This will consist of two class tests.
  
  
• There is also an exam in January which is worth 80% of the final mark for the module.

Approximate Syllabus

• Introduction to knowledge representation and reasoning (KR&R) and basic logics for KR&R:
  
  Introduction;
  Overview of KR&R and its importance for agent systems;
  Review of propositional and first-order predicate logics and their limitations.
  
  
• Modal logics:
  
  Syntax and semantics (Kripke models) of modal logics, model checking, satisfiability checking, theorem proving;
  Logics of knowledge and belief, modelling epistemic puzzles.
  
  
• Non-monotonic reasoning:
  
  Overview of the basics of non-monotonic logics and reasoning with defaults, syntax and semantics of default logic;
  Reasoning through argument.
  
  
• Handling uncertain information through probability and decision theory:
  
  Sample spaces, independence, conditional probability, prior and posterior probabilities, random variables;
  Decision theory for agent systems, Bayesian networks.
  
  

[Back to top]

Lecture Notes

The lecture notes will be printed on a weekly basis as the module proceeds; they can be collected from the Computer Science Helpdesk in the George Holt Building. Subsequent to each lecture electronic versions of the notes will also be posted on VITAL.

[Back to top]