Ontologies and the Semantic Web

Term 1, 201l/2012.

News

• Lab Tutorial, Wednesday Dec 7, 12:00 - 13:00, Room 202.

Office Hour

   After the lectures on Friday (Huxley 421)

Lecture Notes

Full Set of Notes

• formatted as notes [.pdf]
• formatted as slides[.pdf]

Notes by Chapter

• Cover [.pdf]
• Section 1 [.pdf]
• Section 2 [.pdf] (revised)
• Section 3 [.pdf]
• Section 4 [.pdf] (revised)
• Supplement 1 [.pdf]
• Section 5 [.pdf]
• Section 6 [.pdf] (revised)
• Section 7 [.pdf]
• Section 8 [.pdf]
• Section 9 [.pdf]
• Section 10 [.pdf]

Tutorials

• Tutorial 1 (Lab Tutorial) [.pdf]
• Tutorial 2 [.pdf], Model Answers 2 [.pdf]
• Tutorial 3 [.pdf], Model Answers 3 [.pdf] (revised)
• Tutorial 4 [.pdf], Model Answers 4 [.pdf]
• Tutorial 5 [.pdf], Model Answers 5 [.pdf] (revised)
• Tutorial 6 [.pdf] Model Answers 6 [.pdf]
• Tutorial 7 [.pdf], Model Answers 7 [.pdf]
• Tutorial 8 [.pdf], Data Files [.tar.gz]
• Tutorial 9 [.pdf], Model Answers 9 [.pdf]

Coursework

• Coursework Sheet [.pdf]
• Haskell Stub [.hs]
• Prolog Stub [.pl]

General Remarks

• In the Haskell section, read '(Eq r, Eq a) => ...' instead of '(Eq r a) => in the definition of isSatTBox.
• In the Prolog section, read 'nnf' instead of 'reduceF'.
• There is no need that your implementation is optimised in any way. However you may want to think about optimisations: which rules should be applied first? How can we minimize branching?

Test Cases

• satisfiable concepts (empty TBox) [.hs| .pl]
• unsatisfiable concepts (empty TBox) [.hs| .pl]

Literature

• see lecture notes above.

Motivation

With the 21st century dubbed as the information age, readily available access to large pools of knowledge is equally critical for individuals and corporate and government bodies, and indeed for the functioning of modern societies at large. Nowadays, automated tools for processing information increasingly find their way into the mainstream. As stand-alone applications, they allow to reason about knowledge and ontologies expressed in a logical language. One of the prime examples is the NHS-endorsed medical ontology `SNOMED CT' that defines more than 300,000 clinical terms and the corresponding inter-relationships. Embedded into the semantic web, knowledge is represented in precisely the same formalism and extends the human-readable content of a web page with information that can be processed by the machine. In this course, we discuss the family of description logics that provide the underlying tools and techniques for reasoning with, and over, ontologies. We present the basic semantic model, along with the main algorithms and implementation techniques, and students will have the opportunity to implement their own reasoning engine.

Course Description

The course will demonstrate the basic principles of logic-based knowledge representation. The course will be based on description logic as an underlying theoretical foundation, but the goal is to make the course as practical and algorithmic as possible by encouraging the student to implement the basic reasoning tasks themselves, and also use the available tools. The course will discuss the following topics:

• Modelling in description logics: informal examples, ontologies, models and consistency of knowledge bases
• Formal syntax and semantics of ALC and variations, such as versions of DL-Lite used in the semantic web
• Basic reasoning tasks and their relation: concept satisfiability, subsumption, instance checking
• Concept satisfiability in ALC: the case of the empty TBox: tableau-like algorithms and their implementation
• Knowledge bases in ALC and the need for general TBoxes. algorithms for satisfiability over a general TBox and their implementation
• ABoxes and the need for individuals: an extension of ALC with individual variables, reasoning over ABoxes, algorithms and implementation

The aim of the course is to strike a balance between theory and implementation. In class, we will discuss modelling, formal syntax and formal semantics of description logics. The theoretical part of the course then introduces the reasoning tasks, the associated algorithms, together with correctness proofs. In the practical part of the course, students will have the opportunity to implement these reasoning tasks in a prototypical fashion.