Language Technology
Språkteknologi

EDAN20F, 7.5 credits

Valid from: Autumn 2017
Decided by: Professor Thomas Johansson
Date of establishment: 2017-03-22

General Information

Division: Computer Science (LTH)
Course type: Course given jointly for second and third cycle
The course is also given at second-cycle level with course code: EDAN20
Teaching language: English

Aim

In the past 15 years, language technology has considerably matured driven by the massive increase of textual and spoken data and the need to process them automatically. Although there are few systems entirely dedicated to language processing, there are now scores of applications that are to some extent "language-enabled" and embed language processing techniques such as spelling and grammar checkers, information retrieval and extraction, or spoken dialogue systems. This makes the field form a new requirement for the CS engineers. The course introduces theories used in language technology. It attempts to cover the whole field from character encoding and statistical language models to semantics and conversational agents, going through syntax and parsing. It focuses on proven techniques as well as significant industrial or laboratory applications.

Goals

Knowledge and Understanding

For a passing grade the doctoral student must

• Understand the field of language technology and major applications using them
• Know the most important techniques, fundamental algorithms, and most common architectures used in applications
• Create and implement language processing algorithms. Write, interpret, evaluate, and improve them during the programming laboratories.

Competences and Skills

For a passing grade the doctoral student must

• Understand and develop annotation schemes, create and process structured documents
• Understand and write regular expressions and use them in languages like Perl or Java
• Use logic and a logic programming language like Prolog
• Understand and use machine--learning algorithms and statistical techniques
• Develop and evaluate algorithms in major fields of language technology: language models, partial parsing, dependency parsing, and semantic parsing using real data.

Judgement and Approach

For a passing grade the doctoral student must

• Show curiosity, creativity, and problem solving aptitudes
• Show an understanding of industrial and research issues in language technology

Course Contents

An overview of language technology: disciplines, applications, and examples Corpus and word processing: regular expressions, automata, an introduction to Perl, concordances, tokenization, counting words, collocations Morphology and part-of-speech tagging: word morphology, transducers, part-of-speech tagging, Phrase-structure grammars: constituents, trees, DCG rules, unification. Partial parsing: multiword detection, noun group and verb group extraction, information extraction, evaluation Syntax: formalisms, constituency and dependency, functions, parsing, statistical parsing, dependency parsing. Semantics: formal semantics, lambda-calculus, lexical semantics, predicate--argument structures, frame semantics, semantic parsing. Discourse and dialogue: reference and coreference, discourse and rhetoric, discourse relations, parsing discourse relations, dialogue automata, speech acts, multimodality.

Course Literature

Language Processing with Perl and Prolog, Theories, Implementation, and Application. Pierre Nugues, 2014. ISBN 9783642414640.

Instruction Details

Types of instruction: Lectures, laboratory exercises

Examination Details

Examination format: Written exam
Grading scale: Failed, pass
Examiner:

Admission Details

Admission requirements: EDAA01 Programming - Second Course

Course Occasion Information

Contact and Other Information

Course coordinator: Pierre Nugues <pierre.nugues@cs.lth.se>
Web page: http://cs.lth.se/edan20