Semantics and Pragmatics of Natural Language Processing (Level 10) (U02936)

? Credit Points : 10 ? SCQF Level : 10 ? Acronym : INF-4-SPNLP

This module addresses computational approaches to Natural Language (NL) semantics and pragmatics. We will examine several phenomena, including lexical semantics, semantic scope ambiguities, anaphora and presupposition, and resolving semantic ambiguities with both linguistic and non-linguistic information sources. We will examine how notions such as truth, satisfaction and validity can be used to model meaning, and we will explore how semantic representations can be constructed from sentential syntax. We will focus mainly on symbolic techniques, involving logic, inference and theorem proving, but we will also examine how machine learning from corpora can be used to acquire robust semantic interpretations.

The objective is for students to understand some central problems in NL semantics associated with syntax, the lexicon and discourse, as well as methods for handling them, and the trade-offs involved in exploiting hand-crafted, symbolic linguistic models vs. less detailed but more robust statistical models acquired from coprora.

Entry Requirements

? Pre-requisites : Informatics 2A Informatics 2B Successful completion of Year 3 of an Informatics Single or Combined Honours Degree, or year 3 of a Linguistics Single or Combined Honours degree, or equivalent by permission of the school.

? Prohibited combinations : Semantics & Pragmatics of Natural Language Processing (Level 11)

Subject Areas

Home subject area

Language Processing, (School of Informatics, Schedule O)

Delivery Information

? Normal year taken : 4th year

? Delivery Period : Semester 2 (Blocks 3-4)

? Contact Teaching Time : 2 hour(s) per week for 10 weeks

First Class Information

Date	Start	End	Room	Area	Additional Information
08/01/2007	12:10	13:00			Room A9/11, Forrest Hill

All of the following classes

Type	Day	Start	End	Area
Lecture	Monday	12:10	13:00	Central
Lecture	Thursday	12:10	13:00	Central

Summary of Intended Learning Outcomes

It is anticipated that students who successfully complete the course will be able to:
* Describe how theorem proving and model building can be used in assigning interpretations to utterances in NL.
* Evaluate algorithmic and statistical techniques for resolving anaphoric pronouns and noun phrases (NPs).
* Discuss sense-based properties of lexemes and sense-based relations between them, and the effects they have on LT applications.